fix: unify runtime readiness publication and graceful shutdown flow (#3087)

* fix(server): unify runtime readiness and shutdown flow

* refactor(server): decouple readiness shared types

* refactor(server): keep readiness types crate-private

* refactor(server): await protocol shutdown handles

* fix(server): bypass cached startup readiness

.agents/skills/rustfs-release-version-bump/SKILL.md

@@ -0,0 +1,124 @@
+---
+name: rustfs-release-version-bump
+description: "Publish a RustFS alpha/beta/stable release with an auditable flow: confirm target version and scope, update workspace and release assets (including strict rustfs.spec changelog identity/date/version format), run required verification, and finish with commit, push, and GitHub PR creation."
+---
+# RustFS Release Version Bump
+
+Use this skill to publish a RustFS release (alpha, beta, or stable) with a minimal, auditable diff and a complete ship flow (`edit -> verify -> commit -> push -> PR`).
+
+Validated baseline: release pattern used in PR `#2957`.
+
+## Required inputs
+
+- Exact target version, for example `1.0.0-beta.4`.
+- Delivery scope:
+- Local only (`edit/verify`).
+- Local + git (`commit/push`).
+- Full GitHub flow (`commit/push/PR`).
+
+If target version is missing or ambiguous, stop and ask before editing.
+
+## Read before editing
+
+- `AGENTS.md` (root and nearest path-specific files).
+- `.github/pull_request_template.md`.
+- Current branch status and diff against `origin/main`.
+
+## Default release file scope
+
+Treat the following file list as the default checklist for each release bump:
+
+- `Cargo.toml`
+- `Cargo.lock`
+- `README.md`
+- `README_ZH.md`
+- `flake.nix`
+- `helm/rustfs/Chart.yaml`
+- `rustfs.spec`
+
+Only drop a file when the current repository release process clearly no longer requires it.
+
+## Hard release policy
+
+- Docker doc tags use `<version>` (for example `rustfs/rustfs:1.0.0-beta.4`), not `v<version>`.
+- Helm chart version mapping follows `beta.N -> 0.N.0`.
+- `rustfs.spec` `Release` uses prerelease suffix only (for example `beta.4`).
+- Do not change these rules without explicit confirmation.
+
+## Step-by-step workflow
+
+1. Confirm intent and isolate scope
+- Confirm target version string exactly.
+- Confirm whether user requested local-only or full GitHub flow.
+- Inspect current branch and ensure only release-related files are touched for this task.
+
+2. Update workspace versions
+- Bump `[workspace.package].version` in `Cargo.toml`.
+- Bump internal workspace crate dependency versions in `Cargo.toml`.
+- Update `Cargo.lock` so workspace package versions match target version.
+- Re-scan for partial leftovers.
+
+3. Update release assets
+- `README.md` and `README_ZH.md`: update versioned Docker examples to target version.
+- `flake.nix`: update package version to target version.
+- `helm/rustfs/Chart.yaml`:
+- `appVersion` = target version.
+- `version` follows chart mapping rule, for example:
+- `1.0.0-beta.3` -> `0.3.0`
+- `1.0.0-beta.4` -> `0.4.0`
+- `rustfs.spec`:
+- Set `Release` to prerelease suffix (example `beta.4`).
+- Add/update top changelog entry with exact format:
+- `* Thu May 20 2026 houseme <housemecn@gmail.com>`
+- `- Update RPM package to RustFS 1.0.0-beta.4`
+- Changelog identity and time must come from current environment:
+- `git config --get user.name`
+- `git config --get user.email`
+- `date '+%a %b %d %Y'`
+- Changelog version text must match target release version exactly.
+
+4. Verify before shipping
+- Run:
+- `cargo fmt --all`
+- `cargo fmt --all --check`
+- `make pre-commit`
+- If verification passes, run `cargo clean`.
+- If `make pre-commit` fails, return `BLOCKED` with root cause and do not silently widen scope to fix unrelated issues unless user asks.
+
+5. Commit strategy
+- Preferred split when both parts changed:
+- `chore(release): prepare <version>` for `Cargo.toml` and `Cargo.lock`.
+- `chore(release): align release assets for <version>` for docs and packaging files.
+- If user asks for one commit, use one commit.
+- Stage only intended release files; do not include unrelated working tree changes.
+
+6. Push and PR
+- Push branch:
+- `git push -u origin <branch>` (first push), or `git push` (tracking already exists).
+- Create PR with template headings unchanged:
+- `gh pr create --base main --head <branch> --title ... --body-file ...`
+- PR title/body must be English.
+- Use `N/A` for non-applicable template sections.
+- Include verification commands and any `BLOCKED` reason clearly.
+
+## Recommended check commands
+
+- `git status --short --branch`
+- `git diff --name-only origin/main...HEAD`
+- `git diff --stat origin/main...HEAD`
+- `rg -n "<old_version>|<new_version>" Cargo.toml Cargo.lock README.md README_ZH.md flake.nix helm/rustfs/Chart.yaml rustfs.spec`
+- `cargo fmt --all`
+- `cargo fmt --all --check`
+- `make pre-commit`
+- `cargo clean`
+
+## Output contract
+
+When using this skill, always report:
+
+- Target version.
+- Files changed.
+- Any assumptions or uncertainties requiring confirmation.
+- Verification result (`PASSED` or `BLOCKED`) with key evidence.
+- Commit message(s) used.
+- Push status and PR URL when GitHub flow is requested.

.agents/skills/rustfs-release-version-bump/agents/openai.yaml

@@ -0,0 +1,4 @@
+interface:
+  display_name: "RustFS Release Bump"
+  short_description: "Prepare RustFS release branches like PR #2957."
+  default_prompt: "Use $rustfs-release-version-bump to prepare a RustFS release version, ask about any unclear version policy, and finish the commit/push/PR flow."

.agents/skills/security-advisory-lessons/SKILL.md

@@ -0,0 +1,119 @@
+---
+name: security-advisory-lessons
+description: Apply RustFS security lessons distilled from repository GitHub Security Advisories. Use when making or reviewing RustFS code changes, doing security checks, handling PR review for auth/authz, IAM, storage, RPC, logging, CORS, console/browser, encryption, policy, or endpoint changes, and when deciding which security regression tests are required.
+---
+
+# RustFS Security Advisory Lessons
+
+Use this skill as a RustFS-specific security lens before changing or approving code. For the distilled advisory lessons and review patterns, read [advisory-patterns.md](references/advisory-patterns.md).
+
+When currentness matters, fetch the live advisory inventory instead of relying on this skill as a status mirror:
+
+```bash
+gh api repos/rustfs/rustfs/security-advisories --paginate \
+  --jq '.[] | {ghsa_id,state,severity,summary,updated_at}'
+```
+
+Fetch full advisory details only when the live summary suggests a new or changed lesson:
+
+```bash
+gh api repos/rustfs/rustfs/security-advisories/<GHSA_ID>
+```
+
+For the full pattern map, read [advisory-patterns.md](references/advisory-patterns.md).
+
+## Workflow
+
+### 1. Scope the change
+- Identify touched routes, handlers, storage paths, credentials, logs, browser surfaces, CI/release code, and policy checks.
+- Treat these paths as security-sensitive by default: `rustfs/src/admin/`, `rustfs/src/storage/`, `rustfs/src/auth.rs`, `rustfs/src/server/layer.rs`, `crates/iam/`, `crates/policy/`, `crates/credentials/`, `crates/ecstore/src/rpc/`, `crates/rio/`, and console preview/auth code.
+
+### 2. Map to advisory classes
+- Read [advisory-patterns.md](references/advisory-patterns.md) for matching GHSA lessons.
+- Do not rely on advisory titles alone. Confirm whether the issue is authentication, authorization, input validation, storage invariant, browser isolation, logging, or operational hardening.
+
+### 3. Verify fail-closed behavior
+- Check that unauthenticated, wrong-permission, cross-user, cross-bucket, malformed-input, and default-config cases fail explicitly.
+- Prefer exact action/permission checks over broad helper calls or inferred ownership.
+- Confirm lower storage/RPC layers do not bypass checks done in upper layers.
+
+### 4. Require regression evidence
+- For behavior changes, add focused negative tests that reproduce the advisory class.
+- For sensitive fixes, include tests for the bypass form, not only the happy path.
+- If a test is impractical, explain the residual risk and provide a manual verification command.
+
+### 5. Report clearly
+- Lead with concrete findings and file/line evidence.
+- Separate proven vulnerabilities from hardening risks.
+- Avoid exaggerating unauthenticated impact when the code actually rejects unauthenticated requests but allows a low-privileged authenticated bypass.
+
+## Advisory-Derived Guardrails
+
+### Auth and admin authorization
+- Every admin or diagnostic route needs an explicit authn and authz story. Route registration, router whitelist, and handler-level authorization must agree.
+- Match the admin action to the operation exactly. Copy-paste action constants are a known RustFS vulnerability class.
+- Avoid authentication-only helpers for state-changing admin APIs; use `validate_admin_request` or the established equivalent with the right `AdminAction`.
+- Do not assume admin-action `Resource` scoping constrains blast radius unless the policy engine actually enforces resources for that action.
+
+### IAM and service accounts
+- Treat imported IAM payload fields as attacker-controlled: `parent`, `claims`, `accessKey`, `secretKey`, status, policy names, and groups.
+- For service account create/update/import, prove parent ownership or root/admin authority before writing credentials or claims.
+- Do not let `deny_only` or "no explicit deny" become an allow decision that skips required allow checks.
+- Test cross-user list/update/import flows with wrong, correct, self, parent, and root identities.
+
+### S3 copy, multipart, and presigned POST
+- Multipart copy must enforce source `GetObject` and destination `PutObject` semantics equivalent to `CopyObject`, including copy-source and policy conditions.
+- Do not let `CreateMultipartUpload`, `UploadPartCopy`, `CompleteMultipartUpload`, or `AbortMultipartUpload` return success without authorization.
+- Presigned POST policies are server-side contracts. Enforce `content-length-range`, key prefix, exact metadata/content-type, and all signed policy conditions.
+
+### Paths, object keys, and filesystem access
+- Never join untrusted bucket/object/RPC path strings onto filesystem roots without normalization and boundary checks.
+- Reject or safely handle `..`, absolute paths, URL-encoded traversal, platform separators, empty components, and paths that canonicalize outside the intended root.
+- Validate both S3 object-key paths and internode/RPC disk paths; storage helpers can bypass S3 authorization if they trust already-parsed paths.
+
+### Secrets, default credentials, and crypto
+- Do not ship hard-coded shared tokens, HMAC secrets, private keys, or production test keys.
+- Defaults for internode/RPC auth must fail closed for network-reachable deployments or require explicit opt-in with loud warnings.
+- License or token validation must use signatures with embedded public/verifying keys only; do not use private-key decryption as authenticity.
+- Plan key rotation and key IDs when removing exposed keys.
+
+### Logging and debug output
+- Logs must never include access keys beyond safe identifiers, secret keys, session tokens, JWT claims, HMAC secrets, expected signatures, license secrets, or raw response bodies containing credentials.
+- Treat `Debug` implementations, `?value` tracing, merged config dumps, and dependency-level HTTP body logging as leak surfaces.
+- Add log-capture tests or targeted unit tests for redaction wrappers when changing credential structs or response bodies.
+
+### RPC, parsing, and panic safety
+- Treat all RPC payload bytes as attacker-controlled. Replace `unwrap`, `expect`, and panic-prone deserialization with typed errors.
+- Malformed request tests should cover empty bytes, truncated MessagePack/protobuf, invalid enum values, stale timestamps, and invalid signatures.
+- RPC authentication must be independently strong; do not depend on S3 admin credentials unless the fallback is explicit and safe.
+
+### Browser, CORS, and console surfaces
+- Do not reflect arbitrary `Origin` while also allowing credentials. Default CORS should be no CORS unless explicitly configured.
+- Do not render user-controlled object content in a same-origin iframe with console credentials available to JavaScript.
+- Prefer origin separation for object preview/download, `nosniff`, CSP, strict content-type handling, and avoiding durable credentials in `localStorage`.
+- Console license/version-like metadata endpoints should expose only coarse public data unless authenticated, especially subject names and expiration timestamps.
+
+### Profiling, debug, and health endpoints
+- Profiling and debug endpoints are not health checks. They require admin auth, opt-in enablement, rate limiting, and safe responses.
+- Do not return absolute filesystem paths or other deployment layout in unauthenticated or low-privilege responses.
+- Ensure health endpoint allowlists cannot accidentally include expensive diagnostics.
+
+### Trusted proxy and network identity
+- Only honor `X-Forwarded-For` or `X-Real-IP` when the request came from a configured trusted proxy.
+- Direct clients must use the socket peer address for `aws:SourceIp` and policy condition evaluation.
+- Add tests for direct spoofed headers and trusted-proxy headers.
+
+### SSE and storage invariants
+- Encryption metadata is not proof that bytes were encrypted on disk.
+- When touching reader/writer wrappers such as hashing, encryption, compression, or warp readers, verify wrapper order and inspect stored bytes in regression tests.
+- Avoid helper shortcuts that unwrap nested readers and accidentally bypass encryption or integrity layers.
+
+## Review Prompts
+
+Use these prompts while reviewing a diff:
+
+- Could a low-privileged authenticated user reach this path with the wrong action, parent, bucket, or source object?
+- Does a public/default/empty config change security behavior from fail-closed to fail-open?
+- Is any attacker-controlled value later used as a path, policy condition, credential identity, log field, URL, Origin, or response body?
+- Is the same operation implemented in multiple paths, such as `CopyObject` vs `UploadPartCopy`, and do all paths enforce the same security contract?
+- Does the test prove the exploit form is denied, or only that the intended form still works?

.agents/skills/security-advisory-lessons/agents/openai.yaml

@@ -0,0 +1,4 @@
+interface:
+  display_name: "Security Advisory Lessons"
+  short_description: "Apply advisory lessons in reviews."
+  default_prompt: "Review code changes against past RustFS security advisory lessons and report concrete risks, missing tests, and recommended fixes."

.agents/skills/security-advisory-lessons/references/advisory-patterns.md

@@ -0,0 +1,93 @@
+# RustFS Advisory Pattern Map
+
+This file is a lesson map, not an advisory inventory mirror. It keeps durable security patterns distilled from RustFS GitHub Security Advisories.
+
+When current advisory state, severity, URLs, or full text matters, fetch it live:
+
+```bash
+gh api repos/rustfs/rustfs/security-advisories --paginate \
+  --jq '.[] | {ghsa_id,state,severity,summary,updated_at}'
+gh api repos/rustfs/rustfs/security-advisories/<GHSA_ID>
+```
+
+Update this file only when an advisory adds or changes a reusable lesson, affected surface, validation pattern, or regression-test expectation. Do not update it for state-only, URL-only, count-only, or timestamp-only changes.
+
+## Pattern Index
+
+### Admin authorization and route exposure
+
+- `GHSA-pfcq-4gjr-6gjm`: notification target endpoints accepted authenticated users but skipped admin authorization. Lesson: distinguish authn from authz; admin target CRUD must call the operation-specific admin authorization path.
+- `GHSA-mm2q-qcmx-gw4w`: `ListServiceAccount` used `UpdateServiceAccountAdminAction`, while update lacked target ownership checks. Lesson: exact action constants and ownership checks are both required; information disclosure can chain into secret rotation and takeover.
+- `GHSA-vcwh-pff9-64cc`: `ImportIam` checked `ExportIAMAction` for an import/write operation. Lesson: every admin handler must authorize the action it actually performs.
+- `GHSA-jqmc-mg33-v45g` and `GHSA-8784-9m7f-c6p6`: `/profile/cpu` and `/profile/memory` were whitelisted from auth and allowed expensive diagnostics plus path disclosure. Lesson: profiling/debug endpoints need admin auth, opt-in, rate limits, and non-sensitive responses.
+- `GHSA-xp32-gxq2-3v52`: console license metadata endpoint was public and exposed subject and expiration fields. Lesson: management metadata endpoints should require admin auth or return only coarse public status.
+
+### IAM import, service accounts, and privilege boundaries
+
+- `GHSA-566f-q62r-wcr8`: `ImportIam` accepted attacker-controlled service account `parent`, `claims`, `accessKey`, and `secretKey`, enabling persistent backdoor accounts under root. Lesson: imported IAM payloads are untrusted data and must be validated against privilege boundaries.
+- `GHSA-xgr5-qc6w-vcg9`: `deny_only=true` skipped allow checks and let restricted service accounts mint unrestricted children. Lesson: deny-only logic must never become implicit allow for privilege creation.
+- `GHSA-mm2q-qcmx-gw4w`: leaked service account access keys plus update-without-ownership formed an escalation chain. Lesson: service-account identifiers are security-sensitive because update APIs consume them.
+
+### S3 copy, multipart, and upload policy validation
+
+- `GHSA-mx42-j6wv-px98`: `UploadPartCopy` missed source authorization and allowed cross-bucket object exfiltration. Lesson: multipart copy must enforce the same source and destination contract as `CopyObject`.
+- `GHSA-wfxj-ph3v-7mjf`: `UploadPartCopy` checked source and destination independently but missed destination copy-source policy constraints. Lesson: source read and destination write checks are not sufficient when policy constrains allowed copy sources.
+- `GHSA-w5fh-f8xh-5x3p`: presigned POST accepted uploads without enforcing signed policy conditions. Lesson: parse and enforce all POST policy constraints server-side, including size, key prefix, and content type.
+
+### Filesystem paths and object key traversal
+
+- `GHSA-pq29-69jg-9mxc`: RPC `read_file_stream` joined untrusted paths under a volume directory without canonical boundary checks. Lesson: `PathBuf::join` plus length checks are not path security.
+- `GHSA-8r6f-hmq2-28rg`: object keys containing traversal sequences bypassed bucket/object authorization when mapped to filesystem paths. Lesson: reject traversal at object-key parsing and verify final storage paths remain under the expected bucket/key root.
+
+### Secrets, defaults, and cryptographic misuse
+
+- `GHSA-h956-rh7x-ppgj`: gRPC used the hard-coded token `rustfs rpc` on both client and server. Lesson: source-visible shared tokens are authentication bypasses.
+- `GHSA-r5qv-rc46-hv8q`: internode RPC HMAC secret fell back to the public default `rustfsadmin`. Lesson: RPC/internode auth must fail closed instead of silently using public defaults.
+- `GHSA-923g-jp7v-f97f`: license verification embedded a production RSA private key and used private-key decryption as authenticity. Lesson: ship verifying/public keys only and use real signature verification.
+
+### Sensitive logging and debug output
+
+- `GHSA-r54g-49rx-98cr`: STS credentials were logged at info level. Lesson: generated credentials must never be logged in plaintext.
+- `GHSA-8cm2-h255-v749`: debug logs leaked session tokens, secret keys, JWT claims, and raw STS response bodies. Lesson: redaction must cover custom `Debug` implementations and dependency response-body logging.
+- `GHSA-333v-68xh-8mmq`: invalid RPC signature logging included the shared HMAC secret and expected signature. Lesson: error paths often leak secrets; never log raw secrets or derived authenticators.
+
+### RPC input validation and panic safety
+
+- `GHSA-gw2x-q739-qhcr`: malformed gRPC `GetMetrics` payloads reached `unwrap()` on deserialization and caused remote DoS. Lesson: every network/RPC deserialization failure returns an error, not a panic.
+- `GHSA-h956-rh7x-ppgj` and `GHSA-r5qv-rc46-hv8q`: weak RPC auth increased reachability of otherwise internal handlers. Lesson: panic bugs become more severe when internode auth is weak or defaulted.
+
+### Browser, CORS, and console isolation
+
+- `GHSA-v9fg-3cr2-277j`: object preview rendered attacker-controlled HTML in a same-origin iframe, exposing console credentials stored in `localStorage`. Lesson: user content must be origin-isolated from the console and protected with `nosniff`, CSP, and strict content-type handling.
+- `GHSA-x5xv-223c-8vm7`: default CORS reflected arbitrary origins with credentials. Lesson: never combine reflected origins with `Access-Control-Allow-Credentials: true`; default should be fail-closed.
+
+### Trusted proxy and source IP conditions
+
+- `GHSA-fc6g-2gcp-2qrq`: `aws:SourceIp` trusted client-supplied `X-Forwarded-For` or `X-Real-IP`. Lesson: forwarded IP headers are valid only behind configured trusted proxies; direct clients use socket peer IP.
+
+### SSE and on-disk storage invariants
+
+- `GHSA-xrrf-67jm-3c2r`: SSE metadata reported encryption while reader composition bypassed `EncryptReader` and stored plaintext. Lesson: test actual bytes on disk and wrapper order, not only API metadata.
+
+## Useful Search Seeds
+
+Use these targeted searches when a diff touches security-sensitive code:
+
+```bash
+rg -n "validate_admin_request|check_permissions|AdminAction::|deny_only|is_allowed" rustfs crates
+rg -n "UploadPartCopy|upload_part_copy|CompleteMultipart|PostObject|content-length-range|starts-with" rustfs crates
+rg -n "PathBuf::join|canonicalize|\\.\\.|x-forwarded-for|x-real-ip|SourceIp" rustfs crates
+rg -n "DEFAULT_SECRET|DEFAULT_ACCESS|TEST_PRIVATE_KEY|rustfs rpc|RUSTFS_RPC_SECRET" rustfs crates
+rg -n "debug!|trace!|info!|error!|\\?resp|\\?merged_config|session_token|secret_key" rustfs crates
+rg -n "HashReader|EncryptReader|SSE|server-side encryption|Access-Control-Allow-Credentials|Origin" rustfs crates
+```
+
+## Minimum Regression Test Expectations
+
+- Authz fixes: include unauthenticated, valid low-privilege, wrong-action, correct-action, owner, non-owner, and root/admin cases as applicable.
+- IAM fixes: include import/update/list service-account cases with attacker-controlled parent, claims, access key, secret key, and policy.
+- Copy/upload fixes: include cross-bucket, cross-user, source-denied, destination-denied, copy-source-condition, and multipart completion cases.
+- Path fixes: include encoded traversal, absolute path, nested traversal, valid object keys that resemble traversal text but should be rejected, and canonical boundary checks.
+- Logging fixes: assert redacted output for structs and response bodies that may contain credentials.
+- Browser/CORS fixes: assert no credentials on reflected/default origins, correct behavior for explicit allowlists, and no same-origin script execution for previewed object content.
+- SSE fixes: inspect stored bytes and verify API metadata, read-back behavior, and on-disk ciphertext together.

.config/make/help.mak

@@ -4,7 +4,7 @@
 .PHONY: help
 help: ## Shows This Help Menu
 	echo -e "$$HEADER"
-	grep -E '(^[a-zA-Z0-9_-]+:.*?## .*$$)|(^## )' $(MAKEFILE_LIST) | sed 's/^[^:]*://g' | awk 'BEGIN {FS = ":.*?## | #"} ; {printf "${cyan}%-30s${reset} ${white}%s${reset} ${green}%s${reset}\n", $$1, $$2, $$3}' | sed -e 's/\[36m##/\n[32m##/'
+	grep -E '(^[a-zA-Z0-9_-]+:.*?## .*$$)|(^## )' $(MAKEFILE_LIST) | sed 's/^[^:]*://g' | awk 'BEGIN {FS = ":.*?## | #"} /^## / {printf "\n${green}%s${reset}\n", $$0; next} {printf "${cyan}%-30s${reset} ${white}%s${reset} ${green}%s${reset}\n", $$1, $$2, $$3}'
 
 .PHONY: help-build
 help-build: ## Shows RustFS build help

.config/make/lint-fmt.mak

@@ -16,6 +16,11 @@ clippy-check: core-deps ## Run clippy checks
 	cargo clippy --fix --allow-dirty
 	cargo clippy --all-targets --all-features -- -D warnings
 
+.PHONY: unsafe-code-check
+unsafe-code-check: ## Check unsafe_code allowances have SAFETY comments
+	@echo "🔒 Checking unsafe_code allowances..."
+	./scripts/check_unsafe_code_allowances.sh
+
 .PHONY: compilation-check
 compilation-check: core-deps ## Run compilation check
 	@echo "🔨 Running compilation check..."

.config/make/pre-commit.mak

@@ -7,5 +7,5 @@ setup-hooks: ## Set up git hooks
 	@echo "✅ Git hooks setup complete!"
 
 .PHONY: pre-commit
-pre-commit: fmt clippy-check compilation-check test ## Run pre-commit checks
-	@echo "✅ All pre-commit checks passed!"
+pre-commit: fmt unsafe-code-check clippy-check compilation-check test ## Run pre-commit checks
+	@echo "✅ All pre-commit checks passed!"

.config/make/tests.mak

@@ -2,8 +2,13 @@
 
 TEST_THREADS ?= 1
 
+.PHONY: script-tests
+script-tests: ## Run shell script tests
+	@echo "Running script tests..."
+	./scripts/test_build_rustfs_options.sh
+
 .PHONY: test
-test: core-deps test-deps ## Run all tests
+test: core-deps test-deps script-tests ## Run all tests
 	@echo "🧪 Running tests..."
 	@if command -v cargo-nextest >/dev/null 2>&1; then \
 		cargo nextest run --all --exclude e2e_test; \

.docker/compose/README.md

@@ -36,6 +36,64 @@ To start a 4-node cluster for distributed testing:
 docker compose -f .docker/compose/docker-compose.cluster.yaml up -d
 ```
 
+### Script-Based 4-Node Validation (Recommended)
+
+Use the local validation script when you need local-source image build, failover checks,
+and benchmark workflow in one command:
+
+```bash
+# Default mode: WAIT_PROBE_MODE=service
+# This avoids false negatives where /health/ready remains 503 locally
+# while the service path is already available.
+./scripts/run_four_node_cluster_failover_bench.sh
+```
+
+Strict mode is available when you explicitly want `/health/ready == 200` as the gate:
+
+```bash
+WAIT_PROBE_MODE=ready ./scripts/run_four_node_cluster_failover_bench.sh
+```
+
+### Profiling + Trace Validation
+
+The profiling-focused 4-node compose keeps profiling enabled and points RustFS
+to an OTLP/HTTP collector endpoint:
+
+```bash
+docker compose -f .docker/compose/docker-compose.cluster.local-build.profiling-amd64.yml up -d
+```
+
+Important behavior notes:
+
+- `RUSTFS_OBS_ENDPOINT` is the OTLP/HTTP base URL. RustFS automatically sends
+  traces to `/v1/traces`, metrics to `/v1/metrics`, and logs to `/v1/logs`.
+- Startup usually produces logs and metrics first. That does not guarantee
+  visible traces yet.
+- Trace data becomes obvious only after real HTTP/S3/gRPC requests hit RustFS.
+- `RUSTFS_OBS_LOGGER_LEVEL=info` keeps the top-level request span but filters
+  many nested `debug` spans. If Tempo/Jaeger looks sparse, retry with
+  `RUSTFS_OBS_LOGGER_LEVEL=debug` before suspecting the collector.
+
+Minimal trace verification flow:
+
+```bash
+# 1. Start the profiling compose with richer span visibility.
+RUSTFS_OBS_LOGGER_LEVEL=debug \
+docker compose -f .docker/compose/docker-compose.cluster.local-build.profiling-amd64.yml up -d
+
+# 2. Generate real request traffic after startup.
+curl -I http://127.0.0.1:9000/health
+curl -I http://127.0.0.1:9000/health/ready
+
+# 3. Then inspect Tempo or Jaeger.
+# Grafana: http://localhost:3000
+# Jaeger:  http://localhost:16686
+```
+
+If logs and metrics are present but traces are sparse, the most common cause is
+"no real request traffic yet" or "`info` level filtered nested spans", not an
+OTLP routing failure.
+
 ### (Deprecated) Minimal Observability
 
 ```bash

.docker/compose/docker-compose.cluster.local-build.profiling-amd64.yml

@@ -0,0 +1,236 @@
+# Copyright 2024 RustFS Team
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# Profiling-first 4-node local-build compose.
+#
+# Goals:
+# - force linux/amd64 runtime/build on Apple Silicon hosts;
+# - enable RustFS built-in CPU profiling;
+# - keep all tuning knobs host-overridable via env.
+#
+# Observability notes:
+# - `RUSTFS_OBS_ENDPOINT` is the OTLP/HTTP base URL. RustFS appends
+#   `/v1/traces`, `/v1/metrics`, and `/v1/logs` automatically.
+# - Logs and metrics usually appear during startup. Traces mainly appear after
+#   real HTTP/S3/gRPC requests create spans.
+# - `RUSTFS_OBS_LOGGER_LEVEL=info` keeps the top-level request trace span but
+#   filters many `debug`-level nested spans. Use `debug` when validating trace
+#   richness rather than collector reachability.
+
+services:
+  node1:
+    platform: ${RUSTFS_DOCKER_PLATFORM:-linux/amd64}
+    image: ${RUSTFS_IMAGE:-rustfs/rustfs:local-4node}
+    build:
+      context: ../..
+      dockerfile: Dockerfile.source
+    hostname: node1
+    environment:
+      - RUSTFS_VOLUMES=http://node{1...4}:9000/data/rustfs{0...3}
+      - RUSTFS_ADDRESS=:9000
+      - RUSTFS_CONSOLE_ENABLE=true
+      - RUSTFS_ACCESS_KEY=${RUSTFS_ACCESS_KEY:-rustfs-cluster-admin}
+      - RUSTFS_SECRET_KEY=${RUSTFS_SECRET_KEY:-rustfs-cluster-secret}
+      - RUSTFS_OBS_ENDPOINT=${RUSTFS_OBS_ENDPOINT:-http://host.docker.internal:4318}
+      # `info` is enough for startup logs/metrics. Use `debug` if Tempo/Jaeger
+      # should show richer nested spans during request-path verification.
+      - RUSTFS_OBS_LOGGER_LEVEL=${RUSTFS_OBS_LOGGER_LEVEL:-info}
+      - RUSTFS_OBS_PROFILING_ENDPOINT=${RUSTFS_OBS_PROFILING_ENDPOINT:-http://host.docker.internal:4040}
+      - RUSTFS_OBS_PROFILING_EXPORT_ENABLED=${RUSTFS_OBS_PROFILING_EXPORT_ENABLED:-true}
+      - RUSTFS_UNSAFE_BYPASS_DISK_CHECK=${RUSTFS_UNSAFE_BYPASS_DISK_CHECK:-true}
+      - RUSTFS_ENABLE_PROFILING=${RUSTFS_ENABLE_PROFILING:-true}
+      - RUSTFS_PROF_CPU_MODE=${RUSTFS_PROF_CPU_MODE:-continuous}
+      - RUSTFS_PROF_CPU_FREQ=${RUSTFS_PROF_CPU_FREQ:-99}
+      - RUSTFS_PROF_OUTPUT_DIR=${RUSTFS_PROF_OUTPUT_DIR:-/tmp/rustfs-profiles}
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=${RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS:-48}
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=${RUSTFS_OBJECT_IO_BUFFER_SIZE:-262144}
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=${RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD:-6}
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=${RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD:-12}
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=${RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY:-8}
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=${RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE:-8388608}
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=${RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES:-25165824}
+      - RUSTFS_RUNTIME_WORKER_THREADS=${RUSTFS_RUNTIME_WORKER_THREADS:-12}
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=${RUSTFS_RUNTIME_MAX_BLOCKING_THREADS:-512}
+      - RUSTFS_ALLOCATOR_RECLAIM_ENABLED=${RUSTFS_ALLOCATOR_RECLAIM_ENABLED:-false}
+    extra_hosts:
+      - "host.docker.internal:host-gateway"
+    volumes:
+      - node1_data_0:/data/rustfs0
+      - node1_data_1:/data/rustfs1
+      - node1_data_2:/data/rustfs2
+      - node1_data_3:/data/rustfs3
+    ports:
+      - "9000:9000"
+    networks:
+      - rustfs-cluster-net
+
+  node2:
+    platform: ${RUSTFS_DOCKER_PLATFORM:-linux/amd64}
+    image: ${RUSTFS_IMAGE:-rustfs/rustfs:local-4node}
+    build:
+      context: ../..
+      dockerfile: Dockerfile.source
+    hostname: node2
+    environment:
+      - RUSTFS_VOLUMES=http://node{1...4}:9000/data/rustfs{0...3}
+      - RUSTFS_ADDRESS=:9000
+      - RUSTFS_CONSOLE_ENABLE=true
+      - RUSTFS_ACCESS_KEY=${RUSTFS_ACCESS_KEY:-rustfs-cluster-admin}
+      - RUSTFS_SECRET_KEY=${RUSTFS_SECRET_KEY:-rustfs-cluster-secret}
+      - RUSTFS_OBS_ENDPOINT=${RUSTFS_OBS_ENDPOINT:-http://host.docker.internal:4318}
+      # `info` is enough for startup logs/metrics. Use `debug` if Tempo/Jaeger
+      # should show richer nested spans during request-path verification.
+      - RUSTFS_OBS_LOGGER_LEVEL=${RUSTFS_OBS_LOGGER_LEVEL:-info}
+      - RUSTFS_OBS_PROFILING_ENDPOINT=${RUSTFS_OBS_PROFILING_ENDPOINT:-http://host.docker.internal:4040}
+      - RUSTFS_OBS_PROFILING_EXPORT_ENABLED=${RUSTFS_OBS_PROFILING_EXPORT_ENABLED:-true}
+      - RUSTFS_UNSAFE_BYPASS_DISK_CHECK=${RUSTFS_UNSAFE_BYPASS_DISK_CHECK:-true}
+      - RUSTFS_ENABLE_PROFILING=${RUSTFS_ENABLE_PROFILING:-true}
+      - RUSTFS_PROF_CPU_MODE=${RUSTFS_PROF_CPU_MODE:-continuous}
+      - RUSTFS_PROF_CPU_FREQ=${RUSTFS_PROF_CPU_FREQ:-99}
+      - RUSTFS_PROF_OUTPUT_DIR=${RUSTFS_PROF_OUTPUT_DIR:-/tmp/rustfs-profiles}
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=${RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS:-48}
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=${RUSTFS_OBJECT_IO_BUFFER_SIZE:-262144}
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=${RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD:-6}
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=${RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD:-12}
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=${RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY:-8}
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=${RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE:-8388608}
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=${RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES:-25165824}
+      - RUSTFS_RUNTIME_WORKER_THREADS=${RUSTFS_RUNTIME_WORKER_THREADS:-12}
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=${RUSTFS_RUNTIME_MAX_BLOCKING_THREADS:-512}
+      - RUSTFS_ALLOCATOR_RECLAIM_ENABLED=${RUSTFS_ALLOCATOR_RECLAIM_ENABLED:-false}
+    extra_hosts:
+      - "host.docker.internal:host-gateway"
+    volumes:
+      - node2_data_0:/data/rustfs0
+      - node2_data_1:/data/rustfs1
+      - node2_data_2:/data/rustfs2
+      - node2_data_3:/data/rustfs3
+    ports:
+      - "9001:9000"
+    networks:
+      - rustfs-cluster-net
+
+  node3:
+    platform: ${RUSTFS_DOCKER_PLATFORM:-linux/amd64}
+    image: ${RUSTFS_IMAGE:-rustfs/rustfs:local-4node}
+    build:
+      context: ../..
+      dockerfile: Dockerfile.source
+    hostname: node3
+    environment:
+      - RUSTFS_VOLUMES=http://node{1...4}:9000/data/rustfs{0...3}
+      - RUSTFS_ADDRESS=:9000
+      - RUSTFS_CONSOLE_ENABLE=true
+      - RUSTFS_ACCESS_KEY=${RUSTFS_ACCESS_KEY:-rustfs-cluster-admin}
+      - RUSTFS_SECRET_KEY=${RUSTFS_SECRET_KEY:-rustfs-cluster-secret}
+      - RUSTFS_OBS_ENDPOINT=${RUSTFS_OBS_ENDPOINT:-http://host.docker.internal:4318}
+      # `info` is enough for startup logs/metrics. Use `debug` if Tempo/Jaeger
+      # should show richer nested spans during request-path verification.
+      - RUSTFS_OBS_LOGGER_LEVEL=${RUSTFS_OBS_LOGGER_LEVEL:-info}
+      - RUSTFS_OBS_PROFILING_ENDPOINT=${RUSTFS_OBS_PROFILING_ENDPOINT:-http://host.docker.internal:4040}
+      - RUSTFS_OBS_PROFILING_EXPORT_ENABLED=${RUSTFS_OBS_PROFILING_EXPORT_ENABLED:-true}
+      - RUSTFS_UNSAFE_BYPASS_DISK_CHECK=${RUSTFS_UNSAFE_BYPASS_DISK_CHECK:-true}
+      - RUSTFS_ENABLE_PROFILING=${RUSTFS_ENABLE_PROFILING:-true}
+      - RUSTFS_PROF_CPU_MODE=${RUSTFS_PROF_CPU_MODE:-continuous}
+      - RUSTFS_PROF_CPU_FREQ=${RUSTFS_PROF_CPU_FREQ:-99}
+      - RUSTFS_PROF_OUTPUT_DIR=${RUSTFS_PROF_OUTPUT_DIR:-/tmp/rustfs-profiles}
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=${RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS:-48}
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=${RUSTFS_OBJECT_IO_BUFFER_SIZE:-262144}
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=${RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD:-6}
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=${RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD:-12}
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=${RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY:-8}
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=${RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE:-8388608}
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=${RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES:-25165824}
+      - RUSTFS_RUNTIME_WORKER_THREADS=${RUSTFS_RUNTIME_WORKER_THREADS:-12}
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=${RUSTFS_RUNTIME_MAX_BLOCKING_THREADS:-512}
+      - RUSTFS_ALLOCATOR_RECLAIM_ENABLED=${RUSTFS_ALLOCATOR_RECLAIM_ENABLED:-false}
+    extra_hosts:
+      - "host.docker.internal:host-gateway"
+    volumes:
+      - node3_data_0:/data/rustfs0
+      - node3_data_1:/data/rustfs1
+      - node3_data_2:/data/rustfs2
+      - node3_data_3:/data/rustfs3
+    ports:
+      - "9002:9000"
+    networks:
+      - rustfs-cluster-net
+
+  node4:
+    platform: ${RUSTFS_DOCKER_PLATFORM:-linux/amd64}
+    image: ${RUSTFS_IMAGE:-rustfs/rustfs:local-4node}
+    build:
+      context: ../..
+      dockerfile: Dockerfile.source
+    hostname: node4
+    environment:
+      - RUSTFS_VOLUMES=http://node{1...4}:9000/data/rustfs{0...3}
+      - RUSTFS_ADDRESS=:9000
+      - RUSTFS_CONSOLE_ENABLE=true
+      - RUSTFS_ACCESS_KEY=${RUSTFS_ACCESS_KEY:-rustfs-cluster-admin}
+      - RUSTFS_SECRET_KEY=${RUSTFS_SECRET_KEY:-rustfs-cluster-secret}
+      - RUSTFS_OBS_ENDPOINT=${RUSTFS_OBS_ENDPOINT:-http://host.docker.internal:4318}
+      # `info` is enough for startup logs/metrics. Use `debug` if Tempo/Jaeger
+      # should show richer nested spans during request-path verification.
+      - RUSTFS_OBS_LOGGER_LEVEL=${RUSTFS_OBS_LOGGER_LEVEL:-info}
+      - RUSTFS_OBS_PROFILING_ENDPOINT=${RUSTFS_OBS_PROFILING_ENDPOINT:-http://host.docker.internal:4040}
+      - RUSTFS_OBS_PROFILING_EXPORT_ENABLED=${RUSTFS_OBS_PROFILING_EXPORT_ENABLED:-true}
+      - RUSTFS_UNSAFE_BYPASS_DISK_CHECK=${RUSTFS_UNSAFE_BYPASS_DISK_CHECK:-true}
+      - RUSTFS_ENABLE_PROFILING=${RUSTFS_ENABLE_PROFILING:-true}
+      - RUSTFS_PROF_CPU_MODE=${RUSTFS_PROF_CPU_MODE:-continuous}
+      - RUSTFS_PROF_CPU_FREQ=${RUSTFS_PROF_CPU_FREQ:-99}
+      - RUSTFS_PROF_OUTPUT_DIR=${RUSTFS_PROF_OUTPUT_DIR:-/tmp/rustfs-profiles}
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=${RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS:-48}
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=${RUSTFS_OBJECT_IO_BUFFER_SIZE:-262144}
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=${RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD:-6}
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=${RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD:-12}
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=${RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY:-8}
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=${RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE:-8388608}
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=${RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES:-25165824}
+      - RUSTFS_RUNTIME_WORKER_THREADS=${RUSTFS_RUNTIME_WORKER_THREADS:-12}
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=${RUSTFS_RUNTIME_MAX_BLOCKING_THREADS:-512}
+      - RUSTFS_ALLOCATOR_RECLAIM_ENABLED=${RUSTFS_ALLOCATOR_RECLAIM_ENABLED:-false}
+    extra_hosts:
+      - "host.docker.internal:host-gateway"
+    volumes:
+      - node4_data_0:/data/rustfs0
+      - node4_data_1:/data/rustfs1
+      - node4_data_2:/data/rustfs2
+      - node4_data_3:/data/rustfs3
+    ports:
+      - "9003:9000"
+    networks:
+      - rustfs-cluster-net
+
+volumes:
+  node1_data_0:
+  node1_data_1:
+  node1_data_2:
+  node1_data_3:
+  node2_data_0:
+  node2_data_1:
+  node2_data_2:
+  node2_data_3:
+  node3_data_0:
+  node3_data_1:
+  node3_data_2:
+  node3_data_3:
+  node4_data_0:
+  node4_data_1:
+  node4_data_2:
+  node4_data_3:
+
+networks:
+  rustfs-cluster-net:
+    driver: bridge

.docker/compose/docker-compose.cluster.local-build.yml

@@ -0,0 +1,156 @@
+# Copyright 2024 RustFS Team
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+services:
+  node1:
+    image: ${RUSTFS_IMAGE:-rustfs/rustfs:local-4node}
+    build:
+      context: ../..
+      dockerfile: Dockerfile.source
+    hostname: node1
+    environment:
+      - RUSTFS_VOLUMES=${RUSTFS_VOLUMES:-http://node{1...4}:9000/data/rustfs{0...3}}
+      - RUSTFS_ADDRESS=:9000
+      - RUSTFS_CONSOLE_ENABLE=true
+      - RUSTFS_ACCESS_KEY=${RUSTFS_ACCESS_KEY:-rustfs-cluster-admin}
+      - RUSTFS_SECRET_KEY=${RUSTFS_SECRET_KEY:-rustfs-cluster-secret}
+      - RUSTFS_OBS_ENDPOINT=${RUSTFS_OBS_ENDPOINT:-http://host.docker.internal:4318}
+      - RUSTFS_OBS_LOGGER_LEVEL=${RUSTFS_OBS_LOGGER_LEVEL:-info}
+      - RUSTFS_OBS_USE_STDOUT=${RUSTFS_OBS_USE_STDOUT:-false}
+      - RUSTFS_OBS_LOG_STDOUT_ENABLED=${RUSTFS_OBS_LOG_STDOUT_ENABLED:-false}
+      - RUSTFS_ISSUE3031_DIAG_ENABLE=${RUSTFS_ISSUE3031_DIAG_ENABLE:-false}
+      - RUSTFS_UNSAFE_BYPASS_DISK_CHECK=${RUSTFS_UNSAFE_BYPASS_DISK_CHECK:-true}
+    extra_hosts:
+      - "host.docker.internal:host-gateway"
+    volumes:
+      - node1_data_0:/data/rustfs0
+      - node1_data_1:/data/rustfs1
+      - node1_data_2:/data/rustfs2
+      - node1_data_3:/data/rustfs3
+    ports:
+      - "9000:9000"
+    networks:
+      - rustfs-cluster-net
+
+  node2:
+    image: ${RUSTFS_IMAGE:-rustfs/rustfs:local-4node}
+    build:
+      context: ../..
+      dockerfile: Dockerfile.source
+    hostname: node2
+    environment:
+      - RUSTFS_VOLUMES=${RUSTFS_VOLUMES:-http://node{1...4}:9000/data/rustfs{0...3}}
+      - RUSTFS_ADDRESS=:9000
+      - RUSTFS_CONSOLE_ENABLE=true
+      - RUSTFS_ACCESS_KEY=${RUSTFS_ACCESS_KEY:-rustfs-cluster-admin}
+      - RUSTFS_SECRET_KEY=${RUSTFS_SECRET_KEY:-rustfs-cluster-secret}
+      - RUSTFS_OBS_ENDPOINT=${RUSTFS_OBS_ENDPOINT:-http://host.docker.internal:4318}
+      - RUSTFS_OBS_LOGGER_LEVEL=${RUSTFS_OBS_LOGGER_LEVEL:-info}
+      - RUSTFS_OBS_USE_STDOUT=${RUSTFS_OBS_USE_STDOUT:-false}
+      - RUSTFS_OBS_LOG_STDOUT_ENABLED=${RUSTFS_OBS_LOG_STDOUT_ENABLED:-false}
+      - RUSTFS_ISSUE3031_DIAG_ENABLE=${RUSTFS_ISSUE3031_DIAG_ENABLE:-false}
+      - RUSTFS_UNSAFE_BYPASS_DISK_CHECK=${RUSTFS_UNSAFE_BYPASS_DISK_CHECK:-true}
+    extra_hosts:
+      - "host.docker.internal:host-gateway"
+    volumes:
+      - node2_data_0:/data/rustfs0
+      - node2_data_1:/data/rustfs1
+      - node2_data_2:/data/rustfs2
+      - node2_data_3:/data/rustfs3
+    ports:
+      - "9001:9000"
+    networks:
+      - rustfs-cluster-net
+
+  node3:
+    image: ${RUSTFS_IMAGE:-rustfs/rustfs:local-4node}
+    build:
+      context: ../..
+      dockerfile: Dockerfile.source
+    hostname: node3
+    environment:
+      - RUSTFS_VOLUMES=${RUSTFS_VOLUMES:-http://node{1...4}:9000/data/rustfs{0...3}}
+      - RUSTFS_ADDRESS=:9000
+      - RUSTFS_CONSOLE_ENABLE=true
+      - RUSTFS_ACCESS_KEY=${RUSTFS_ACCESS_KEY:-rustfs-cluster-admin}
+      - RUSTFS_SECRET_KEY=${RUSTFS_SECRET_KEY:-rustfs-cluster-secret}
+      - RUSTFS_OBS_ENDPOINT=${RUSTFS_OBS_ENDPOINT:-http://host.docker.internal:4318}
+      - RUSTFS_OBS_LOGGER_LEVEL=${RUSTFS_OBS_LOGGER_LEVEL:-info}
+      - RUSTFS_OBS_USE_STDOUT=${RUSTFS_OBS_USE_STDOUT:-false}
+      - RUSTFS_OBS_LOG_STDOUT_ENABLED=${RUSTFS_OBS_LOG_STDOUT_ENABLED:-false}
+      - RUSTFS_ISSUE3031_DIAG_ENABLE=${RUSTFS_ISSUE3031_DIAG_ENABLE:-false}
+      - RUSTFS_UNSAFE_BYPASS_DISK_CHECK=${RUSTFS_UNSAFE_BYPASS_DISK_CHECK:-true}
+    extra_hosts:
+      - "host.docker.internal:host-gateway"
+    volumes:
+      - node3_data_0:/data/rustfs0
+      - node3_data_1:/data/rustfs1
+      - node3_data_2:/data/rustfs2
+      - node3_data_3:/data/rustfs3
+    ports:
+      - "9002:9000"
+    networks:
+      - rustfs-cluster-net
+
+  node4:
+    image: ${RUSTFS_IMAGE:-rustfs/rustfs:local-4node}
+    build:
+      context: ../..
+      dockerfile: Dockerfile.source
+    hostname: node4
+    environment:
+      - RUSTFS_VOLUMES=${RUSTFS_VOLUMES:-http://node{1...4}:9000/data/rustfs{0...3}}
+      - RUSTFS_ADDRESS=:9000
+      - RUSTFS_CONSOLE_ENABLE=true
+      - RUSTFS_ACCESS_KEY=${RUSTFS_ACCESS_KEY:-rustfs-cluster-admin}
+      - RUSTFS_SECRET_KEY=${RUSTFS_SECRET_KEY:-rustfs-cluster-secret}
+      - RUSTFS_OBS_ENDPOINT=${RUSTFS_OBS_ENDPOINT:-http://host.docker.internal:4318}
+      - RUSTFS_OBS_LOGGER_LEVEL=${RUSTFS_OBS_LOGGER_LEVEL:-info}
+      - RUSTFS_OBS_USE_STDOUT=${RUSTFS_OBS_USE_STDOUT:-false}
+      - RUSTFS_OBS_LOG_STDOUT_ENABLED=${RUSTFS_OBS_LOG_STDOUT_ENABLED:-false}
+      - RUSTFS_ISSUE3031_DIAG_ENABLE=${RUSTFS_ISSUE3031_DIAG_ENABLE:-false}
+      - RUSTFS_UNSAFE_BYPASS_DISK_CHECK=${RUSTFS_UNSAFE_BYPASS_DISK_CHECK:-true}
+    extra_hosts:
+      - "host.docker.internal:host-gateway"
+    volumes:
+      - node4_data_0:/data/rustfs0
+      - node4_data_1:/data/rustfs1
+      - node4_data_2:/data/rustfs2
+      - node4_data_3:/data/rustfs3
+    ports:
+      - "9003:9000"
+    networks:
+      - rustfs-cluster-net
+
+volumes:
+  node1_data_0:
+  node1_data_1:
+  node1_data_2:
+  node1_data_3:
+  node2_data_0:
+  node2_data_1:
+  node2_data_2:
+  node2_data_3:
+  node3_data_0:
+  node3_data_1:
+  node3_data_2:
+  node3_data_3:
+  node4_data_0:
+  node4_data_1:
+  node4_data_2:
+  node4_data_3:
+
+networks:
+  rustfs-cluster-net:
+    driver: bridge

.docker/compose/docker-compose.cluster.perf-round1-linux-32c64g-latency-stable.override.yml

@@ -0,0 +1,56 @@
+services:
+  node1:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=56
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=131072
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=10
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=20
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=10
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=8388608
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=25165824
+      - RUSTFS_RUNTIME_WORKER_THREADS=16
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=768
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=420
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=50
+
+  node2:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=56
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=131072
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=10
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=20
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=10
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=8388608
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=25165824
+      - RUSTFS_RUNTIME_WORKER_THREADS=16
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=768
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=420
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=50
+
+  node3:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=56
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=131072
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=10
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=20
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=10
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=8388608
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=25165824
+      - RUSTFS_RUNTIME_WORKER_THREADS=16
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=768
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=420
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=50
+
+  node4:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=56
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=131072
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=10
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=20
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=10
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=8388608
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=25165824
+      - RUSTFS_RUNTIME_WORKER_THREADS=16
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=768
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=420
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=50

.docker/compose/docker-compose.cluster.perf-round1-linux-32c64g-throughput-max.override.yml

@@ -0,0 +1,56 @@
+services:
+  node1:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=96
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=524288
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=16
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=32
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=24
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=33554432
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=50331648
+      - RUSTFS_RUNTIME_WORKER_THREADS=24
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=1536
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=600
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=80
+
+  node2:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=96
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=524288
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=16
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=32
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=24
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=33554432
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=50331648
+      - RUSTFS_RUNTIME_WORKER_THREADS=24
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=1536
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=600
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=80
+
+  node3:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=96
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=524288
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=16
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=32
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=24
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=33554432
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=50331648
+      - RUSTFS_RUNTIME_WORKER_THREADS=24
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=1536
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=600
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=80
+
+  node4:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=96
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=524288
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=16
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=32
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=24
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=33554432
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=50331648
+      - RUSTFS_RUNTIME_WORKER_THREADS=24
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=1536
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=600
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=80

.docker/compose/docker-compose.cluster.perf-round1-linux-32c64g.override.yml

@@ -0,0 +1,56 @@
+services:
+  node1:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=64
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=262144
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=12
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=24
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=16
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=16777216
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=33554432
+      - RUSTFS_RUNTIME_WORKER_THREADS=20
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=1024
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=300
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=60
+
+  node2:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=64
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=262144
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=12
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=24
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=16
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=16777216
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=33554432
+      - RUSTFS_RUNTIME_WORKER_THREADS=20
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=1024
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=300
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=60
+
+  node3:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=64
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=262144
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=12
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=24
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=16
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=16777216
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=33554432
+      - RUSTFS_RUNTIME_WORKER_THREADS=20
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=1024
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=300
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=60
+
+  node4:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=64
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=262144
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=12
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=24
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=16
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=16777216
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=33554432
+      - RUSTFS_RUNTIME_WORKER_THREADS=20
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=1024
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=300
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=60

.docker/compose/docker-compose.cluster.perf-round1-linux.override.yml

@@ -0,0 +1,56 @@
+services:
+  node1:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=48
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=262144
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=8
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=16
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=12
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=8388608
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=25165824
+      - RUSTFS_RUNTIME_WORKER_THREADS=12
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=512
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=300
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=40
+
+  node2:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=48
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=262144
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=8
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=16
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=12
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=8388608
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=25165824
+      - RUSTFS_RUNTIME_WORKER_THREADS=12
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=512
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=300
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=40
+
+  node3:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=48
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=262144
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=8
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=16
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=12
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=8388608
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=25165824
+      - RUSTFS_RUNTIME_WORKER_THREADS=12
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=512
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=300
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=40
+
+  node4:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=48
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=262144
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=8
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=16
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=12
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=8388608
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=25165824
+      - RUSTFS_RUNTIME_WORKER_THREADS=12
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=512
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=300
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=40

.docker/compose/docker-compose.cluster.perf-round1.override.yml

@@ -0,0 +1,56 @@
+services:
+  node1:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=32
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=262144
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=6
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=12
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=8
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=8388608
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=16777216
+      - RUSTFS_RUNTIME_WORKER_THREADS=6
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=256
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=300
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=30
+
+  node2:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=32
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=262144
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=6
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=12
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=8
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=8388608
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=16777216
+      - RUSTFS_RUNTIME_WORKER_THREADS=6
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=256
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=300
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=30
+
+  node3:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=32
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=262144
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=6
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=12
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=8
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=8388608
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=16777216
+      - RUSTFS_RUNTIME_WORKER_THREADS=6
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=256
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=300
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=30
+
+  node4:
+    environment:
+      - RUSTFS_OBJECT_MAX_CONCURRENT_DISK_READS=32
+      - RUSTFS_OBJECT_IO_BUFFER_SIZE=262144
+      - RUSTFS_OBJECT_MEDIUM_CONCURRENCY_THRESHOLD=6
+      - RUSTFS_OBJECT_HIGH_CONCURRENCY_THRESHOLD=12
+      - RUSTFS_OBJECT_IO_RANDOM_READAHEAD_DISABLE_CONCURRENCY=8
+      - RUSTFS_OBJECT_DUPLEX_BUFFER_SIZE=8388608
+      - RUSTFS_ERASURE_ENCODE_MAX_INFLIGHT_BYTES=16777216
+      - RUSTFS_RUNTIME_WORKER_THREADS=6
+      - RUSTFS_RUNTIME_MAX_BLOCKING_THREADS=256
+      - RUSTFS_CAPACITY_SCHEDULED_INTERVAL=300
+      - RUSTFS_CAPACITY_WRITE_FREQUENCY_THRESHOLD=30

.docker/compose/docker-compose.cluster.yaml

@@ -21,8 +21,8 @@ services:
       - RUSTFS_VOLUMES=http://node{0...3}:9000/data/rustfs{0...3}
       - RUSTFS_ADDRESS=0.0.0.0:9000
       - RUSTFS_CONSOLE_ENABLE=true
-      - RUSTFS_ACCESS_KEY=rustfsadmin
-      - RUSTFS_SECRET_KEY=rustfsadmin
+      - RUSTFS_ACCESS_KEY=rustfs-cluster-admin
+      - RUSTFS_SECRET_KEY=rustfs-cluster-secret
     platform: linux/amd64
     ports:
       - "9000:9000" # Map port 9001 of the host to port 9000 of the container
@@ -38,8 +38,8 @@ services:
       - RUSTFS_VOLUMES=http://node{0...3}:9000/data/rustfs{0...3}
       - RUSTFS_ADDRESS=0.0.0.0:9000
       - RUSTFS_CONSOLE_ENABLE=true
-      - RUSTFS_ACCESS_KEY=rustfsadmin
-      - RUSTFS_SECRET_KEY=rustfsadmin
+      - RUSTFS_ACCESS_KEY=rustfs-cluster-admin
+      - RUSTFS_SECRET_KEY=rustfs-cluster-secret
     platform: linux/amd64
     ports:
       - "9001:9000" # Map port 9002 of the host to port 9000 of the container
@@ -55,8 +55,8 @@ services:
       - RUSTFS_VOLUMES=http://node{0...3}:9000/data/rustfs{0...3}
       - RUSTFS_ADDRESS=0.0.0.0:9000
       - RUSTFS_CONSOLE_ENABLE=true
-      - RUSTFS_ACCESS_KEY=rustfsadmin
-      - RUSTFS_SECRET_KEY=rustfsadmin
+      - RUSTFS_ACCESS_KEY=rustfs-cluster-admin
+      - RUSTFS_SECRET_KEY=rustfs-cluster-secret
     platform: linux/amd64
     ports:
       - "9002:9000" # Map port 9003 of the host to port 9000 of the container
@@ -72,8 +72,8 @@ services:
       - RUSTFS_VOLUMES=http://node{0...3}:9000/data/rustfs{0...3}
       - RUSTFS_ADDRESS=0.0.0.0:9000
       - RUSTFS_CONSOLE_ENABLE=true
-      - RUSTFS_ACCESS_KEY=rustfsadmin
-      - RUSTFS_SECRET_KEY=rustfsadmin
+      - RUSTFS_ACCESS_KEY=rustfs-cluster-admin
+      - RUSTFS_SECRET_KEY=rustfs-cluster-secret
     platform: linux/amd64
     ports:
       - "9003:9000" # Map port 9004 of the host to port 9000 of the container

.docker/compose/docker-compose.observability.yaml

@@ -26,7 +26,7 @@ services:
     restart: "no"
 
   tempo:
-    image: grafana/tempo:latest
+    image: grafana/tempo:2.10.5
     user: "10001"
     command: [ "-config.file=/etc/tempo.yaml" ]
     volumes:
@@ -36,9 +36,19 @@ services:
       - "3200:3200" # tempo
       - "4317"      # otlp grpc
       - "4318"      # otlp http
+      - "7946"      # memberlist
     restart: unless-stopped
     networks:
       - rustfs-network
+    depends_on:
+      tempo-init:
+        condition: service_completed_successfully
+    healthcheck:
+      test: [ "CMD", "/tempo", "-version" ]
+      interval: 10s
+      timeout: 5s
+      retries: 3
+      start_period: 15s
 
   otel-collector:
     image: otel/opentelemetry-collector-contrib:latest
@@ -61,6 +71,12 @@ services:
       - jaeger
       - prometheus
       - loki
+    restart: unless-stopped
+    healthcheck:
+      test: [ "CMD", "/otelcol-contrib", "--version" ]
+      interval: 10s
+      timeout: 5s
+      retries: 3
 
   jaeger:
     image: jaegertracing/jaeger:latest
@@ -72,12 +88,23 @@ services:
       - BADGER_DIRECTORY_KEY=/badger/key
       - COLLECTOR_OTLP_ENABLED=true
     volumes:
+      - ../../.docker/observability/jaeger.yaml:/etc/jaeger/config.yml:ro
       - jaeger-data:/badger
     ports:
       - "16686:16686" # Web UI
       - "14269:14269" # Admin/Metrics
+      - "4317"        # otlp grpc
+      - "4318"        # otlp http
+    command: [ "--config", "/etc/jaeger/config.yml" ]
     networks:
       - rustfs-network
+    restart: unless-stopped
+    healthcheck:
+      test: [ "CMD", "wget", "--spider", "-q", "http://localhost:14269" ]
+      interval: 10s
+      timeout: 5s
+      retries: 3
+      start_period: 15s
 
   prometheus:
     image: prom/prometheus:latest
@@ -85,6 +112,7 @@ services:
       - TZ=Asia/Shanghai
     volumes:
       - ../../.docker/observability/prometheus.yml:/etc/prometheus/prometheus.yml:ro
+      - ../../.docker/observability/prometheus-rules:/etc/prometheus/rules:ro
       - prometheus-data:/prometheus
     ports:
       - "9090:9090"
@@ -94,23 +122,45 @@ services:
       - '--web.enable-remote-write-receiver'
       - '--enable-feature=promql-experimental-functions'
       - '--storage.tsdb.path=/prometheus'
-      - '--web.console.libraries=/usr/share/prometheus/console_libraries'
-      - '--web.console.templates=/usr/share/prometheus/consoles'
+      - '--storage.tsdb.retention.time=30d'
     networks:
       - rustfs-network
+    restart: unless-stopped
+    healthcheck:
+      test: [ "CMD", "wget", "--spider", "-q", "http://localhost:9090/-/healthy" ]
+      interval: 10s
+      timeout: 5s
+      retries: 3
 
   loki:
     image: grafana/loki:latest
     environment:
       - TZ=Asia/Shanghai
     volumes:
-      - ../../.docker/observability/loki.yaml:/etc/loki/local-config.yaml:ro
+      - ../../.docker/observability/loki.yaml:/etc/loki/loki.yaml:ro
       - loki-data:/loki
     ports:
       - "3100:3100"
-    command: -config.file=/etc/loki/local-config.yaml
+    command: -config.file=/etc/loki/loki.yaml
     networks:
       - rustfs-network
+    restart: unless-stopped
+    healthcheck:
+      test: [ "CMD", "wget", "--spider", "-q", "http://localhost:3100/ready" ]
+      interval: 15s
+      timeout: 10s
+      retries: 5
+      start_period: 60s
+
+  pyroscope:
+    image: grafana/pyroscope:latest
+    ports:
+      - "4040:4040"
+    command:
+      - -self-profiling.disable-push=true
+    networks:
+      - rustfs-network
+    restart: unless-stopped
 
   grafana:
     image: grafana/grafana:latest
@@ -127,10 +177,17 @@ services:
     volumes:
       - ../../.docker/observability/grafana/provisioning:/etc/grafana/provisioning:ro
       - ../../.docker/observability/grafana/dashboards:/var/lib/grafana/dashboards:ro
+      - grafana-data:/var/lib/grafana
     depends_on:
       - prometheus
       - tempo
       - loki
+    restart: unless-stopped
+    healthcheck:
+      test: [ "CMD", "wget", "--spider", "-q", "http://localhost:3000/api/health" ]
+      interval: 10s
+      timeout: 5s
+      retries: 3
 
   # --- RustFS Cluster ---
 
@@ -215,6 +272,7 @@ volumes:
   tempo-data:
   loki-data:
   jaeger-data:
+  grafana-data:
 
 networks:
   rustfs-network:

.docker/nginx/nginx.conf

@@ -74,6 +74,17 @@ http {
     #     ssl_certificate /etc/nginx/ssl/server.crt;
     #     ssl_certificate_key /etc/nginx/ssl/server.key;
     #
+    #     # Restrict to modern TLS versions and ciphers. Operators copying this
+    #     # example must keep at least these directives — without them, nginx
+    #     # may negotiate older protocol versions that have known weaknesses.
+    #     ssl_protocols TLSv1.2 TLSv1.3;
+    #     ssl_prefer_server_ciphers on;
+    #     ssl_ciphers 'ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-AES256-GCM-SHA384:ECDHE-RSA-AES256-GCM-SHA384:ECDHE-ECDSA-CHACHA20-POLY1305:ECDHE-RSA-CHACHA20-POLY1305';
+    #     ssl_session_timeout 1d;
+    #     ssl_session_cache shared:SSL:10m;
+    #     ssl_session_tickets off;
+    #     # add_header Strict-Transport-Security "max-age=63072000" always;
+    #
     #     location / {
     #         proxy_pass http://rustfs:9000;
     #         ...

.docker/observability/README.md

@@ -13,6 +13,9 @@ The stack is composed of the following best-in-class open-source components:
 - **Jaeger** (v1.59.0): Distributed tracing system (configured as a secondary UI/storage).
 - **OpenTelemetry Collector** (v0.104.0): A vendor-agnostic implementation for receiving, processing, and exporting telemetry data.
 
+By default, this stack uses Tempo in single-binary mode and does not require Kafka/Redpanda.
+If you want the Kafka-backed HA Tempo path, use `docker-compose-example-for-rustfs.yml` together with `docker-compose-tempo-ha-override.yml`.
+
 ## Architecture
 
 1. **Telemetry Collection**: Applications send OTLP (OpenTelemetry Protocol) data (Metrics, Logs, Traces) to the **OpenTelemetry Collector**.
@@ -46,6 +49,15 @@ Run the following command to start the entire stack:
 docker compose up -d
 ```
 
+### High Availability Tempo
+
+The default `docker-compose.yml` is the single-node stack.
+If you need the Kafka-backed HA Tempo configuration, start it with:
+
+```bash
+docker compose -f docker-compose-example-for-rustfs.yml -f docker-compose-tempo-ha-override.yml up -d
+```
+
 ### Access Dashboards
 
 | Service        | URL                                              | Credentials       | Description                    |
@@ -78,6 +90,54 @@ docker compose down -v
 - **Grafana**: Dashboards and datasources are provisioned from the `grafana/` directory.
 - **Collector**: Edit `otel-collector-config.yaml` to modify pipelines, processors, or exporters.
 
+### Verifying RustFS Traces
+
+When RustFS points `RUSTFS_OBS_ENDPOINT` at this stack, treat the value as the
+OTLP/HTTP base URL, for example:
+
+```bash
+export RUSTFS_OBS_ENDPOINT=http://host.docker.internal:4318
+```
+
+RustFS automatically expands that base URL to:
+
+- `/v1/traces`
+- `/v1/metrics`
+- `/v1/logs`
+
+Important behavior notes:
+
+- Logs and metrics usually appear during startup, so seeing those two signals
+  first is expected.
+- Visible trace data usually requires real HTTP/S3/gRPC request traffic after
+  startup, because request-path spans are created on demand.
+- `RUSTFS_OBS_LOGGER_LEVEL=info` keeps the top-level request span but filters
+  many nested `debug` spans. If Tempo or Jaeger looks sparse, retry with
+  `RUSTFS_OBS_LOGGER_LEVEL=debug` before suspecting collector or Tempo issues.
+
+Minimal validation flow:
+
+```bash
+# 1. Start this observability stack.
+docker compose up -d
+
+# 2. Start RustFS with OTLP/HTTP export and richer span visibility.
+export RUSTFS_OBS_ENDPOINT=http://host.docker.internal:4318
+export RUSTFS_OBS_LOGGER_LEVEL=debug
+
+# 3. Generate real request traffic.
+curl -I http://127.0.0.1:9000/health
+curl -I http://127.0.0.1:9000/health/ready
+
+# 4. Inspect Grafana or Jaeger.
+# Grafana: http://localhost:3000
+# Jaeger:  http://localhost:16686
+```
+
+If logs and metrics are present but traces are sparse, the most common cause is
+"no real request traffic yet" or "`info` level filtered nested spans", not an
+OTLP routing failure.
+
 ## Troubleshooting
 
 - **Service Health**: Check the health of services using `docker compose ps`.

.docker/observability/README_ZH.md

@@ -13,6 +13,9 @@
 - **Jaeger** (v1.59.0): 分布式追踪系统（配置为辅助 UI/存储）。
 - **OpenTelemetry Collector** (v0.104.0): 接收、处理和导出遥测数据的供应商无关实现。
 
+默认情况下，这套技术栈使用 Tempo 单二进制模式，不依赖 Kafka/Redpanda。
+如果需要基于 Kafka 的 HA Tempo 路径，请使用 `docker-compose-example-for-rustfs.yml` 配合 `docker-compose-tempo-ha-override.yml`。
+
 ## 架构
 
 1.  **遥测收集**: 应用程序将 OTLP (OpenTelemetry Protocol) 数据（指标、日志、追踪）发送到 **OpenTelemetry Collector**。
@@ -46,6 +49,15 @@
 docker compose up -d
 ```
 
+### Tempo 高可用模式
+
+默认的 `docker-compose.yml` 对应单机栈。
+如果需要基于 Kafka 的 HA Tempo 配置，请使用：
+
+```bash
+docker compose -f docker-compose-example-for-rustfs.yml -f docker-compose-tempo-ha-override.yml up -d
+```
+
 ### 访问仪表盘
 
 | 服务 | URL | 凭据 | 描述 |
@@ -78,6 +90,50 @@ docker compose down -v
 -   **Grafana**: 仪表盘和数据源从 `grafana/` 目录预置。
 -   **Collector**: 编辑 `otel-collector-config.yaml` 以修改管道、处理器或导出器。
 
+### 验证 RustFS Trace
+
+当 RustFS 将 `RUSTFS_OBS_ENDPOINT` 指向这套技术栈时，应将该值视为
+OTLP/HTTP 的基础 URL，例如：
+
+```bash
+export RUSTFS_OBS_ENDPOINT=http://host.docker.internal:4318
+```
+
+RustFS 会自动在该基础 URL 后补全：
+
+- `/v1/traces`
+- `/v1/metrics`
+- `/v1/logs`
+
+需要注意：
+
+- 启动阶段通常会先看到日志和指标，因此“先有日志/指标、后有 trace”是正常现象。
+- 可见的 trace 数据通常依赖启动后的真实 HTTP/S3/gRPC 请求流量，因为请求路径上的 span 是按需创建的。
+- `RUSTFS_OBS_LOGGER_LEVEL=info` 会保留顶层请求 span，但会过滤掉很多 `debug` 级别的嵌套 span。
+  如果 Tempo 或 Jaeger 中的 trace 看起来很稀疏，建议先改成 `RUSTFS_OBS_LOGGER_LEVEL=debug`，再判断是否是 collector 或 Tempo 问题。
+
+最小验证流程：
+
+```bash
+# 1. 启动本目录下的可观测性技术栈。
+docker compose up -d
+
+# 2. 以 OTLP/HTTP 导出方式启动 RustFS，并提高 span 可见性。
+export RUSTFS_OBS_ENDPOINT=http://host.docker.internal:4318
+export RUSTFS_OBS_LOGGER_LEVEL=debug
+
+# 3. 产生真实请求流量。
+curl -I http://127.0.0.1:9000/health
+curl -I http://127.0.0.1:9000/health/ready
+
+# 4. 到 Grafana 或 Jaeger 中检查。
+# Grafana: http://localhost:3000
+# Jaeger:  http://localhost:16686
+```
+
+如果日志和指标已经正常，但 trace 仍然稀疏，最常见的原因通常是
+“还没有真实请求流量”或“`info` 级别过滤了嵌套 span”，而不是 OTLP 路由失败。
+
 ## 故障排除
 
 -   **服务健康**: 使用 `docker compose ps` 检查服务健康状况。

.docker/observability/docker-compose-example-for-rustfs.yml

@@ -85,10 +85,8 @@ services:
     networks:
       - otel-network
     restart: unless-stopped
-    depends_on:
-      - redpanda
     healthcheck:
-      test: [ "CMD", "wget", "--spider", "-q", "http://localhost:3200/ready" ]
+      test: [ "CMD", "/tempo", "-version" ]
       interval: 10s
       timeout: 5s
       retries: 3
@@ -205,7 +203,7 @@ services:
       - prometheus
       - loki
     healthcheck:
-      test: [ "CMD", "wget", "--spider", "-q", "http://localhost:13133" ]
+      test: [ "CMD", "/otelcol-contrib", "--version" ]
       interval: 10s
       timeout: 5s
       retries: 3

.docker/observability/docker-compose.yml

@@ -16,10 +16,8 @@ services:
 
   # --- Tracing ---
   tempo:
-    image: grafana/tempo:2.10.3
+    image: grafana/tempo:2.10.5
     container_name: tempo
-    depends_on:
-      - redpanda
     command: [ "-config.file=/etc/tempo.yaml" ]
     volumes:
       - ./tempo.yaml:/etc/tempo.yaml:ro
@@ -33,31 +31,11 @@ services:
       - otel-network
     restart: unless-stopped
     healthcheck:
-      test: [ "CMD", "wget", "--spider", "-q", "http://localhost:3200/ready" ]
+      test: [ "CMD", "/tempo", "-version" ]
       interval: 10s
       timeout: 5s
       retries: 3
       start_period: 15s
-  redpanda:
-    image: redpandadata/redpanda:latest
-    ports:
-      - "9092:9092" # Kafka API for clients
-    command: >
-      redpanda start --overprovisioned
-      --mode=dev-container
-      --kafka-addr=PLAINTEXT://0.0.0.0:9092
-      --advertise-kafka-addr=PLAINTEXT://redpanda:9092
-
-  redpanda-console:
-    image: docker.redpanda.com/redpandadata/console:latest
-    environment:
-      - CONFIG_FILEPATH=/etc/redpanda/redpanda-console-config.yaml
-    volumes:
-      - ./redpanda-console.yaml:/etc/redpanda/redpanda-console-config.yaml
-    ports:
-      - "8080:8080"
-    depends_on:
-      - redpanda
 
   vulture:
     image: grafana/tempo-vulture:latest
@@ -106,6 +84,7 @@ services:
     container_name: prometheus
     volumes:
       - ./prometheus.yml:/etc/prometheus/prometheus.yml:ro
+      - ./prometheus-rules:/etc/prometheus/rules:ro
       - prometheus-data:/prometheus
     ports:
       - "9090:9090"
@@ -168,7 +147,7 @@ services:
       - prometheus
       - loki
     healthcheck:
-      test: [ "CMD", "wget", "--spider", "-q", "http://localhost:13133" ]
+      test: [ "CMD", "/otelcol-contrib", "--version" ]
       interval: 10s
       timeout: 5s
       retries: 3

.docker/observability/grafana/provisioning/datasources.yaml

@@ -1,97 +0,0 @@
-# Copyright 2024 RustFS Team
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#     http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-apiVersion: 1
-
-datasources:
-  - name: Prometheus
-    type: prometheus
-    uid: prometheus
-    access: proxy
-    orgId: 1
-    url: http://prometheus:9090
-    isDefault: true
-    version: 1
-    editable: false
-    jsonData:
-      httpMethod: GET
-      exemplarTraceIdDestinations:
-        - name: trace_id
-          datasourceUid: tempo
-
-  - name: Tempo
-    type: tempo
-    uid: tempo
-    access: proxy
-    orgId: 1
-    url: http://tempo:3200
-    isDefault: false
-    version: 1
-    editable: false
-    jsonData:
-      httpMethod: GET
-      serviceMap:
-        datasourceUid: prometheus
-      tracesToLogs:
-        datasourceUid: loki
-        tags: [ 'job', 'instance', 'pod', 'namespace', 'service.name' ]
-        mappedTags: [ { key: 'service.name', value: 'app' } ]
-        spanStartTimeShift: '1s'
-        spanEndTimeShift: '-1s'
-        filterByTraceID: true
-        filterBySpanID: false
-      tracesToMetrics:
-        datasourceUid: prometheus
-        tags: [ { key: 'service.name' }, { key: 'job' } ]
-        queries:
-          - name: 'Service-Level Latency'
-            query: 'sum(rate(traces_spanmetrics_latency_bucket{$$__tags}[5m])) by (le)'
-          - name: 'Service-Level Calls'
-            query: 'sum(rate(traces_spanmetrics_calls_total{$$__tags}[5m]))'
-          - name: 'Service-Level Errors'
-            query: 'sum(rate(traces_spanmetrics_calls_total{status_code="ERROR", $$__tags}[5m]))'
-      nodeGraph:
-        enabled: true
-
-  - name: Loki
-    type: loki
-    uid: loki
-    orgId: 1
-    url: http://loki:3100
-    isDefault: false
-    version: 1
-    editable: false
-    jsonData:
-      derivedFields:
-        - datasourceUid: tempo
-          matcherRegex: 'trace_id=(\w+)'
-          name: 'TraceID'
-          url: '$${__value.raw}'
-
-  - name: Jaeger
-    type: jaeger
-    uid: jaeger
-    url: http://jaeger:16686
-    access: proxy
-    isDefault: false
-    editable: false
-    jsonData:
-      tracesToLogs:
-        datasourceUid: loki
-        tags: [ 'job', 'instance', 'pod', 'namespace', 'service.name' ]
-        mappedTags: [ { key: 'service.name', value: 'app' } ]
-        spanStartTimeShift: '1s'
-        spanEndTimeShift: '-1s'
-        filterByTraceID: true
-        filterBySpanID: false

.docker/observability/prometheus-rules/rustfs-dashboard.yml

@@ -0,0 +1,53 @@
+groups:
+  - name: rustfs-dashboard
+    interval: 30s
+    rules:
+      - record: rustfs:http_server_requests:rate5m
+        expr: sum by (job) (rate(rustfs_http_server_requests_total[5m]))
+
+      - record: rustfs:http_server_request_duration_seconds:p50_5m
+        expr: histogram_quantile(0.50, sum by (le, job) (rate(rustfs_http_server_request_duration_seconds_bucket[5m])))
+      - record: rustfs:http_server_request_duration_seconds:p95_5m
+        expr: histogram_quantile(0.95, sum by (le, job) (rate(rustfs_http_server_request_duration_seconds_bucket[5m])))
+      - record: rustfs:http_server_request_duration_seconds:p99_5m
+        expr: histogram_quantile(0.99, sum by (le, job) (rate(rustfs_http_server_request_duration_seconds_bucket[5m])))
+
+      - record: rustfs:http_server_response_body_size_bytes:p50_5m
+        expr: histogram_quantile(0.50, sum by (le, job) (rate(rustfs_http_server_response_body_size_bytes_bucket[5m])))
+      - record: rustfs:http_server_response_body_size_bytes:p95_5m
+        expr: histogram_quantile(0.95, sum by (le, job) (rate(rustfs_http_server_response_body_size_bytes_bucket[5m])))
+      - record: rustfs:http_server_response_body_size_bytes:p99_5m
+        expr: histogram_quantile(0.99, sum by (le, job) (rate(rustfs_http_server_response_body_size_bytes_bucket[5m])))
+
+      - record: rustfs:log_cleaner_runs:rate15m
+        expr: sum by (job) (rate(rustfs_log_cleaner_runs_total[15m]))
+      - record: rustfs:log_cleaner_failure_ratio:rate5m
+        expr: sum by (job) (rate(rustfs_log_cleaner_run_failures_total[5m])) / clamp_min(sum by (job) (rate(rustfs_log_cleaner_runs_total[5m])), 1e-9)
+      - record: rustfs:log_cleaner_rotation_failure_ratio:rate5m
+        expr: sum by (job) (rate(rustfs_log_cleaner_rotation_failures_total[5m])) / clamp_min(sum by (job) (rate(rustfs_log_cleaner_rotation_total[5m])), 1e-9)
+      - record: rustfs:log_cleaner_rotation_duration_seconds:p95_5m
+        expr: histogram_quantile(0.95, sum by (le, job) (rate(rustfs_log_cleaner_rotation_duration_seconds_bucket[5m])))
+      - record: rustfs:log_cleaner_compress_duration_seconds:p95_5m
+        expr: histogram_quantile(0.95, sum by (le, job) (rate(rustfs_log_cleaner_compress_duration_seconds_bucket[5m])))
+
+      - record: rustfs:scanner_objects_scanned:rate5m
+        expr: sum by (job) (rate(rustfs_scanner_objects_scanned_total[5m]))
+      - record: rustfs:scanner_directories_scanned:rate5m
+        expr: sum by (job) (rate(rustfs_scanner_directories_scanned_total[5m]))
+      - record: rustfs:scanner_buckets_scanned:rate5m
+        expr: sum by (job) (rate(rustfs_scanner_buckets_scanned_total[5m]))
+      - record: rustfs:scanner_cycles_success:rate5m
+        expr: sum by (job) (rate(rustfs_scanner_cycles_total{result="success"}[5m]))
+
+      - record: rustfs:log_chain_op_event_mismatch:rate5m
+        expr: sum by (job) (rate(rustfs_log_chain_op_event_mismatch_total[5m]))
+
+      - alert: RustFSLogChainOpEventMismatchDetected
+        expr: rustfs:log_chain_op_event_mismatch:rate5m > 0
+        for: 10m
+        labels:
+          severity: warning
+          component: s3-log-chain
+        annotations:
+          summary: "RustFS log-chain op/event mismatch detected"
+          description: "job={{ $labels.job }} has non-zero rustfs_log_chain_op_event_mismatch_total rate for more than 10m. Check s3 op/event mapping changes."

.docker/observability/prometheus.yml

@@ -19,6 +19,9 @@ global:
     cluster: 'rustfs-dev'  # Label to identify the cluster
     replica: '1'  # Replica identifier
 
+rule_files:
+  - /etc/prometheus/rules/*.yml
+
 scrape_configs:
   - job_name: 'otel-collector'
     static_configs:

.docker/observability/tempo.yaml

@@ -12,7 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-partition_ring_live_store: true
 stream_over_http_enabled: true
 
 server:
@@ -33,33 +32,17 @@ distributor:
           endpoint: "0.0.0.0:4317"
         http:
           endpoint: "0.0.0.0:4318"
-  #log_received_spans:
-  #    enabled: true
-  #  log_discarded_spans:
-  #    enabled: true
 
-backend_scheduler:
-  provider:
-    compaction:
-      compaction:
-        block_retention: 1h
-
-backend_worker:
-  backend_scheduler_addr: localhost:3200
-  compaction:
-    block_retention: 1h
-  ring:
-    kvstore:
-      store: memberlist
-
-querier:
-  query_live_store: true
+ingester:
+  max_block_duration: 5m
 
 metrics_generator:
   registry:
     external_labels:
       source: tempo
       cluster: docker-compose
+  traces_storage:
+    path: /var/tempo/generator/traces
   storage:
     path: /var/tempo/generator/wal
     remote_write:
@@ -85,15 +68,5 @@ overrides:
       processors: [ "span-metrics", "service-graphs", "local-blocks" ]
       generate_native_histograms: both
 
-ingest:
-  enabled: true
-  kafka:
-    address: redpanda:9092
-    topic: tempo-ingest
-
-block_builder:
-  consume_cycle_duration: 30s
-
 usage_report:
   reporting_enabled: false
-

.docker/test/issue-2715/docker-compose-test.yml

@@ -0,0 +1,52 @@
+services:
+  rustfs:
+    image: rustfs/rustfs:1.0.0-alpha.99-glibc
+    container_name: rustfs-issue-2715-test
+    security_opt:
+      - "no-new-privileges:true"
+    ports:
+      - "19000:9000"
+      - "19001:9001"
+    environment:
+      - RUSTFS_VOLUMES=/data/rustfs{0...8}
+      - RUSTFS_ADDRESS=0.0.0.0:9000
+      - RUSTFS_CONSOLE_ADDRESS=0.0.0.0:9001
+      - RUSTFS_CONSOLE_ENABLE=true
+      - RUSTFS_CORS_ALLOWED_ORIGINS=*
+      - RUSTFS_CONSOLE_CORS_ALLOWED_ORIGINS=*
+      - RUSTFS_ACCESS_KEY=admin
+      - RUSTFS_SECRET_KEY=admin
+      - RUSTFS_OBS_LOGGER_LEVEL=info
+      - RUSTFS_OBS_ENDPOINT=http://otel-collector:4318
+      - RUSTFS_OBS_PROFILING_ENDPOINT=http://pyroscope:4040
+      - RUSTFS_STORAGE_CLASS_STANDARD=EC:2
+      - RUSTFS_STORAGE_CLASS_RRS=EC:1
+      - RUSTFS_UNSAFE_BYPASS_DISK_CHECK=true
+      - RUSTFS_OBS_LOG_DIRECTORY=/opt/rustfs/logs
+    extra_hosts:
+      - "otel-collector:host-gateway"
+      - "pyroscope:host-gateway"
+    volumes:
+      - ./deploy/data/issue-2715/rustfs0:/data/rustfs0
+      - ./deploy/data/issue-2715/rustfs1:/data/rustfs1
+      - ./deploy/data/issue-2715/rustfs2:/data/rustfs2
+      - ./deploy/data/issue-2715/rustfs3:/data/rustfs3
+      - ./deploy/data/issue-2715/rustfs4:/data/rustfs4
+      - ./deploy/data/issue-2715/rustfs5:/data/rustfs5
+      - ./deploy/data/issue-2715/rustfs6:/data/rustfs6
+      - ./deploy/data/issue-2715/rustfs7:/data/rustfs7
+      - ./deploy/data/issue-2715/rustfs8:/data/rustfs8
+      - ./deploy/logs/issue-2715:/opt/rustfs/logs
+    restart: unless-stopped
+    healthcheck:
+      test:
+        [
+          "CMD",
+          "sh",
+          "-c",
+          "curl -f http://127.0.0.1:9000/health && curl -f http://127.0.0.1:9001/rustfs/console/health"
+        ]
+      interval: 30s
+      timeout: 10s
+      retries: 3
+      start_period: 40s

.docker/test/issues-2815/.gitignore

@@ -0,0 +1 @@
+data/

.docker/test/issues-2815/README.md

@@ -0,0 +1,106 @@
+# Issue 2815 Local Docker Verification
+
+## Purpose
+
+This directory contains the local distributed Docker verification assets used to validate issue `#2815` against the current source build.
+
+The target behavior is:
+
+- 4-node distributed cluster starts successfully
+- `/health/ready` becomes reachable on each node
+- logs no longer contain `storage_info failed: Io error: wrong msgpack marker FixArray(1)`
+- internode RPC authentication succeeds with an explicit non-default RPC secret
+
+## Files
+
+- `docker-compose.yml`: 4-node distributed cluster using a locally built image
+
+## Data Directories
+
+Create the bind-mount directories before `docker compose up`:
+
+```bash
+mkdir -p .docker/test/issues-2815/data/rustfs{1..4}-disk{0..3}
+```
+
+## Build
+
+Apple Silicon / arm64 host:
+
+```bash
+docker build --platform linux/arm64 -f Dockerfile.source -t rustfs-issue-2815-local .
+```
+
+If you intentionally want amd64 emulation:
+
+```bash
+docker build --platform linux/amd64 -f Dockerfile.source -t rustfs-issue-2815-local .
+```
+
+## Run
+
+```bash
+docker compose -f .docker/test/issues-2815/docker-compose.yml up -d
+```
+
+If the image platform is not `linux/arm64`, align compose explicitly:
+
+```bash
+RUSTFS_DOCKER_PLATFORM=linux/amd64 docker compose -f .docker/test/issues-2815/docker-compose.yml up -d
+```
+
+## Health Checks
+
+Container-level healthcheck is now included and probes:
+
+```bash
+curl -fsS http://127.0.0.1:9000/health
+```
+
+Manual checks:
+
+```bash
+curl -i http://127.0.0.1:9101/health/ready
+curl -i http://127.0.0.1:9102/health/ready
+curl -i http://127.0.0.1:9103/health/ready
+curl -i http://127.0.0.1:9104/health/ready
+```
+
+## RPC Secret Requirement
+
+The current source build no longer reproduces the original `FixArray(1)` decode error from issue `#2815`.
+
+Earlier local Docker attempts failed during erasure bootstrap with:
+
+```text
+No valid auth token
+store init failed to load formats after 10 retries: erasure read quorum
+```
+
+Root cause:
+
+- RPC authentication rejects the default secret `rustfsadmin`
+- distributed local Docker validation therefore needs an explicit non-default secret
+
+This compose now sets both:
+
+- `RUSTFS_SECRET_KEY=issue-2815-secret`
+- `RUSTFS_RPC_SECRET=issue-2815-rpc-secret`
+
+With those values in place, the current 4-node local Docker cluster reaches healthy state and `/health/ready` returns `200`.
+
+In other words:
+
+- `RUSTFS_ACCESS_KEY` may still be `rustfsadmin` for local service credentials if desired
+- `RUSTFS_SECRET_KEY` can still be used for service credentials
+- but RPC authentication must not resolve to the default secret value `rustfsadmin`
+- if `RUSTFS_RPC_SECRET` is unset, the code falls back to `RUSTFS_SECRET_KEY`
+- so at least one of them must provide a non-default shared secret for internode RPC signing
+
+## Suggested Debug Commands
+
+```bash
+docker compose -f .docker/test/issues-2815/docker-compose.yml ps
+docker compose -f .docker/test/issues-2815/docker-compose.yml logs --no-color --tail=200
+docker compose -f .docker/test/issues-2815/docker-compose.yml down -v
+```

.docker/test/issues-2815/docker-compose.yml

@@ -0,0 +1,120 @@
+services:
+  rustfs1:
+    image: rustfs-issue-2815-local
+    platform: ${RUSTFS_DOCKER_PLATFORM:-linux/arm64}
+    hostname: rustfs1
+    container_name: rustfs-issue-2815-rustfs1
+    environment:
+      RUSTFS_ADDRESS: "0.0.0.0:9000"
+      RUSTFS_ACCESS_KEY: "rustfsadmin"
+      RUSTFS_SECRET_KEY: "issue-2815-secret"
+      RUSTFS_RPC_SECRET: "issue-2815-rpc-secret"
+      RUSTFS_CONSOLE_ENABLE: "false"
+      RUST_LOG: "info"
+      RUSTFS_UNSAFE_BYPASS_DISK_CHECK: "true"
+      RUSTFS_VOLUMES: "http://rustfs{1...4}:9000/data/rustfs{0...3}"
+    volumes:
+      - ./data/rustfs1-disk0:/data/rustfs0
+      - ./data/rustfs1-disk1:/data/rustfs1
+      - ./data/rustfs1-disk2:/data/rustfs2
+      - ./data/rustfs1-disk3:/data/rustfs3
+    networks: [rustfs-issue-2815-net]
+    ports:
+      - "9101:9000"
+    healthcheck:
+      test: ["CMD", "sh", "-c", "curl -fsS http://127.0.0.1:9000/health || exit 1"]
+      interval: 15s
+      timeout: 5s
+      retries: 8
+      start_period: 30s
+
+  rustfs2:
+    image: rustfs-issue-2815-local
+    platform: ${RUSTFS_DOCKER_PLATFORM:-linux/arm64}
+    hostname: rustfs2
+    container_name: rustfs-issue-2815-rustfs2
+    environment:
+      RUSTFS_ADDRESS: "0.0.0.0:9000"
+      RUSTFS_ACCESS_KEY: "rustfsadmin"
+      RUSTFS_SECRET_KEY: "issue-2815-secret"
+      RUSTFS_RPC_SECRET: "issue-2815-rpc-secret"
+      RUSTFS_CONSOLE_ENABLE: "false"
+      RUST_LOG: "info"
+      RUSTFS_UNSAFE_BYPASS_DISK_CHECK: "true"
+      RUSTFS_VOLUMES: "http://rustfs{1...4}:9000/data/rustfs{0...3}"
+    volumes:
+      - ./data/rustfs2-disk0:/data/rustfs0
+      - ./data/rustfs2-disk1:/data/rustfs1
+      - ./data/rustfs2-disk2:/data/rustfs2
+      - ./data/rustfs2-disk3:/data/rustfs3
+    networks: [rustfs-issue-2815-net]
+    ports:
+      - "9102:9000"
+    healthcheck:
+      test: ["CMD", "sh", "-c", "curl -fsS http://127.0.0.1:9000/health || exit 1"]
+      interval: 15s
+      timeout: 5s
+      retries: 8
+      start_period: 30s
+
+  rustfs3:
+    image: rustfs-issue-2815-local
+    platform: ${RUSTFS_DOCKER_PLATFORM:-linux/arm64}
+    hostname: rustfs3
+    container_name: rustfs-issue-2815-rustfs3
+    environment:
+      RUSTFS_ADDRESS: "0.0.0.0:9000"
+      RUSTFS_ACCESS_KEY: "rustfsadmin"
+      RUSTFS_SECRET_KEY: "issue-2815-secret"
+      RUSTFS_RPC_SECRET: "issue-2815-rpc-secret"
+      RUSTFS_CONSOLE_ENABLE: "false"
+      RUST_LOG: "info"
+      RUSTFS_UNSAFE_BYPASS_DISK_CHECK: "true"
+      RUSTFS_VOLUMES: "http://rustfs{1...4}:9000/data/rustfs{0...3}"
+    volumes:
+      - ./data/rustfs3-disk0:/data/rustfs0
+      - ./data/rustfs3-disk1:/data/rustfs1
+      - ./data/rustfs3-disk2:/data/rustfs2
+      - ./data/rustfs3-disk3:/data/rustfs3
+    networks: [rustfs-issue-2815-net]
+    ports:
+      - "9103:9000"
+    healthcheck:
+      test: ["CMD", "sh", "-c", "curl -fsS http://127.0.0.1:9000/health || exit 1"]
+      interval: 15s
+      timeout: 5s
+      retries: 8
+      start_period: 30s
+
+  rustfs4:
+    image: rustfs-issue-2815-local
+    platform: ${RUSTFS_DOCKER_PLATFORM:-linux/arm64}
+    hostname: rustfs4
+    container_name: rustfs-issue-2815-rustfs4
+    environment:
+      RUSTFS_ADDRESS: "0.0.0.0:9000"
+      RUSTFS_ACCESS_KEY: "rustfsadmin"
+      RUSTFS_SECRET_KEY: "issue-2815-secret"
+      RUSTFS_RPC_SECRET: "issue-2815-rpc-secret"
+      RUSTFS_CONSOLE_ENABLE: "false"
+      RUST_LOG: "info"
+      RUSTFS_UNSAFE_BYPASS_DISK_CHECK: "true"
+      RUSTFS_VOLUMES: "http://rustfs{1...4}:9000/data/rustfs{0...3}"
+    volumes:
+      - ./data/rustfs4-disk0:/data/rustfs0
+      - ./data/rustfs4-disk1:/data/rustfs1
+      - ./data/rustfs4-disk2:/data/rustfs2
+      - ./data/rustfs4-disk3:/data/rustfs3
+    networks: [rustfs-issue-2815-net]
+    ports:
+      - "9104:9000"
+    healthcheck:
+      test: ["CMD", "sh", "-c", "curl -fsS http://127.0.0.1:9000/health || exit 1"]
+      interval: 15s
+      timeout: 5s
+      retries: 8
+      start_period: 30s
+
+networks:
+  rustfs-issue-2815-net:
+    name: rustfs-issue-2815-net

.github/actions/setup/action.yml

@@ -53,18 +53,19 @@ runs:
           musl-tools \
           build-essential \
           pkg-config \
-          libssl-dev
+          libssl-dev \
+          protobuf-compiler
 
     - name: Install protoc
-      uses: arduino/setup-protoc@v3
+      uses: rustfs/setup-protoc@v3.0.1
       with:
-        version: "33.1"
-        repo-token: ${{ inputs.github-token }}
+        version: "34.1"
+        repo-token: ${{ github.token }}
 
     - name: Install flatc
-      uses: Nugine/setup-flatc@v1.2.4
+      uses: Nugine/setup-flatc@v1
       with:
-        version: "25.9.23"
+        version: "25.12.19"
 
     - name: Install Rust toolchain
       uses: dtolnay/rust-toolchain@stable

.github/dependabot.yml

@@ -27,9 +27,8 @@ updates:
       timezone: "Asia/Shanghai"
       time: "08:00"
     assignees:
-      - "heihutu"
-    reviewers:
       - "houseme"
+    reviewers:
       - "overtrue"
       - "majinghe"
     ignore:
@@ -39,6 +38,8 @@ updates:
         versions: [ "0.23.x" ]
       - dependency-name: "ratelimit"
         versions: [ "1.x" ]
+      - dependency-name: "ratelimit"
+        versions: [ "2.x" ]
     groups:
       s3s:
         update-types:

.github/pull_request_template.md

@@ -2,35 +2,34 @@
 Pull Request Template for RustFS
 -->
 
-## Type of Change
-- [ ] New Feature
-- [ ] Bug Fix
-- [ ] Documentation
-- [ ] Performance Improvement
-- [ ] Test/CI
-- [ ] Refactor
-- [ ] Other:
-
 ## Related Issues
-<!-- List related Issue numbers, e.g. #123 -->
+<!--
+List related issues, e.g. Fixes #123.
+Use N/A when there is no related issue.
+-->
 
 ## Summary of Changes
-<!-- Briefly describe the main changes and motivation for this PR -->
+<!--
+Briefly explain what changed and why reviewers should accept it.
+Focus on behavior, compatibility, and review-relevant context.
+-->
 
-## Checklist
-- [ ] I have read and followed the [CONTRIBUTING.md](CONTRIBUTING.md) guidelines
-- [ ] Passed `make pre-commit`
-- [ ] Added/updated necessary tests
-- [ ] Documentation updated (if needed)
-- [ ] CI/CD passed (if applicable)
+## Verification
+<!--
+List the commands or checks you ran, for example:
+- `make pre-commit`
+
+Use N/A only when verification is not applicable.
+-->
 
 ## Impact
-- [ ] Breaking change (compatibility)
-- [ ] Requires doc/config/deployment update
-- [ ] Other impact:
+<!--
+Describe user-facing, compatibility, API, deployment, configuration, or
+documentation impact. Use N/A when there is no expected impact.
+-->
 
 ## Additional Notes
-<!-- Any extra information for reviewers -->
+<!-- Any extra information for reviewers, or N/A. -->
 
 ---
 

.github/workflows/audit.yml

@@ -75,7 +75,7 @@ jobs:
         uses: actions/checkout@v6
 
       - name: Dependency Review
-        uses: actions/dependency-review-action@v4
+        uses: actions/dependency-review-action@v5
         with:
           fail-on-severity: moderate
-          comment-summary-in-pr: true
+          comment-summary-in-pr: always

.github/workflows/build.yml

@@ -169,7 +169,7 @@ jobs:
             {"target_id":"linux-x86_64-gnu","os":"ubicloud-standard-2","target":"x86_64-unknown-linux-gnu","cross":false,"platform":"linux","rustflags":""},
             {"target_id":"linux-aarch64-gnu","os":"ubicloud-standard-2","target":"aarch64-unknown-linux-gnu","cross":true,"platform":"linux","rustflags":""},
             {"target_id":"macos-aarch64","os":"macos-latest","target":"aarch64-apple-darwin","cross":false,"platform":"macos","rustflags":""},
-            {"target_id":"macos-x86_64","os":"macos-latest","target":"x86_64-apple-darwin","cross":false,"platform":"macos","rustflags":""},
+            {"target_id":"macos-x86_64","os":"macos-15-intel","target":"x86_64-apple-darwin","cross":false,"platform":"macos","rustflags":""},
             {"target_id":"windows-x86_64","os":"windows-latest","target":"x86_64-pc-windows-msvc","cross":false,"platform":"windows","rustflags":""}
           ]}'
 
@@ -234,8 +234,7 @@ jobs:
         run: |
           mkdir -p ./rustfs/static
           if [[ "${{ matrix.platform }}" == "windows" ]]; then
-            curl.exe -L "https://dl.rustfs.com/artifacts/console/rustfs-console-latest.zip" -o console.zip --retry 3 --retry-delay 5 --max-time 300
-            if [[ $? -eq 0 ]]; then
+            if curl.exe --fail -L "https://dl.rustfs.com/artifacts/console/rustfs-console-latest.zip" -o console.zip --retry 3 --retry-delay 5 --max-time 300; then
               unzip -o console.zip -d ./rustfs/static
               rm console.zip
             else
@@ -244,9 +243,8 @@ jobs:
             fi
           else
             chmod +w ./rustfs/static/LICENSE || true
-            curl -L "https://dl.rustfs.com/artifacts/console/rustfs-console-latest.zip" \
-              -o console.zip --retry 3 --retry-delay 5 --max-time 300
-            if [[ $? -eq 0 ]]; then
+            if curl --fail -L "https://dl.rustfs.com/artifacts/console/rustfs-console-latest.zip" \
+              -o console.zip --retry 3 --retry-delay 5 --max-time 300; then
               unzip -o console.zip -d ./rustfs/static
               rm console.zip
             else

.github/workflows/ci.yml

@@ -130,6 +130,9 @@ jobs:
       - name: Check code formatting
         run: cargo fmt --all --check
 
+      - name: Check unsafe code allowances
+        run: ./scripts/check_unsafe_code_allowances.sh
+
       - name: Run clippy lints
         run: cargo clippy --all-targets --all-features -- -D warnings
 
@@ -201,7 +204,7 @@ jobs:
         with:
           tool: s3s-e2e
           git: https://github.com/s3s-project/s3s.git
-          rev: 4a04a670cf41274d9be9ab65dc36f4aa3f92fbad
+          rev: 62cb4a71dd759a6ec56b64c4c42fcc183a2c6a52
 
       - name: Run end-to-end tests
         run: |

.github/workflows/docker.yml

@@ -163,11 +163,10 @@ jobs:
                 if [[ "$version" == *"alpha"* ]] || [[ "$version" == *"beta"* ]] || [[ "$version" == *"rc"* ]]; then
                   build_type="prerelease"
                   is_prerelease=true
-                  # TODO: Temporary change - currently allows alpha versions to also create latest tags
-                  # After the version is stable, you need to remove the following line and restore the original logic (latest is created only for stable versions)
-                  if [[ "$version" == *"alpha"* ]]; then
+                  # Current policy: create latest tags for stable releases and selected prereleases (alpha/beta).
+                  if [[ "$version" == *"alpha"* ]] || [[ "$version" == *"beta"* ]]; then
                     create_latest=true
-                     echo "🧪 Building Docker image for prerelease: $version (temporarily allowing creation of latest tag)"
+                    echo "🧪 Building Docker image for prerelease: $version (creating latest tag)"
                   else
                     echo "🧪 Building Docker image for prerelease: $version"
                   fi
@@ -216,11 +215,10 @@ jobs:
               v*alpha*|v*beta*|v*rc*|*alpha*|*beta*|*rc*)
                 build_type="prerelease"
                 is_prerelease=true
-                # TODO: Temporary change - currently allows alpha versions to also create latest tags
-                # After the version is stable, you need to remove the if block below and restore the original logic.
-                if [[ "$input_version" == *"alpha"* ]]; then
+                # Current policy: create latest tags for stable releases and selected prereleases (alpha/beta).
+                if [[ "$version" == *"alpha"* ]] || [[ "$version" == *"beta"* ]]; then
                   create_latest=true
-                   echo "🧪 Building with prerelease version: $input_version (temporarily allowing creation of latest tag)"
+                  echo "🧪 Building with prerelease version: $input_version (creating latest tag)"
                 else
                   echo "🧪 Building with prerelease version: $input_version"
                 fi
@@ -351,9 +349,7 @@ jobs:
 
           # Add channel tags for prereleases and latest for stable
           if [[ "$CREATE_LATEST" == "true" ]]; then
-            # TODO: Temporary change - the current alpha version will also create the latest tag
-            # After the version is stabilized, the logic here remains unchanged, but the upstream CREATE_LATEST setting needs to be restored.
-            # Stable release (and temporary alpha versions)
+            # Create latest tags for stable releases and selected prereleases when CREATE_LATEST=true.
             TAGS="$TAGS,${{ env.REGISTRY_DOCKERHUB }}:latest${VARIANT_SUFFIX},${{ env.REGISTRY_GHCR }}:latest${VARIANT_SUFFIX},${{ env.REGISTRY_QUAY }}:latest${VARIANT_SUFFIX}"
           elif [[ "$BUILD_TYPE" == "prerelease" ]]; then
             # Prerelease channel tags (alpha, beta, rc)
@@ -450,10 +446,9 @@ jobs:
             "prerelease")
               echo "🧪 Prerelease Docker image has been built with ${VERSION} tags"
               echo "⚠️  This is a prerelease image - use with caution"
-              # TODO: Temporary change - alpha versions currently create the latest tag
-              # After the version is stable, you need to restore the following prompt information
-              if [[ "$VERSION" == *"alpha"* ]] && [[ "$CREATE_LATEST" == "true" ]]; then
-                echo "🏷️  Latest tag has been created for alpha version (temporary measures)"
+              # Create latest tags for stable releases and selected prereleases when CREATE_LATEST=true.
+              if [[ "$CREATE_LATEST" == "true" ]]; then
+                echo "🏷️  Latest tag has been created for prerelease: $VERSION"
               else
                 echo "🚫 Latest tag NOT created for prerelease"
               fi

.github/workflows/e2e-s3tests.yml

@@ -40,8 +40,8 @@ on:
 
 env:
   # main user
-  S3_ACCESS_KEY: rustfsadmin
-  S3_SECRET_KEY: rustfsadmin
+  S3_ACCESS_KEY: rustfs-ci-admin
+  S3_SECRET_KEY: rustfs-ci-secret
   # alt user (must be different from main for many s3-tests)
   S3_ALT_ACCESS_KEY: rustfsalt
   S3_ALT_SECRET_KEY: rustfsalt

.github/workflows/helm-package.yml

@@ -18,41 +18,78 @@ on:
   workflow_run:
     workflows: [ "Build and Release" ]
     types: [ completed ]
+  workflow_dispatch:
+    inputs:
+      version:
+        description: "Release version to publish, e.g. 1.0.0-beta.1 or v1.0.0-beta.1"
+        required: true
+        default: "1.0.0-beta.1"
+        type: string
 
 permissions:
   contents: read
 
-env:
-  new_version: ${{ github.event.workflow_run.head_branch }}
-
 jobs:
   build-helm-package:
     runs-on: ubicloud-standard-2
-    # Only run on successful builds triggered by tag pushes (version format: x.y.z or x.y.z-suffix)
     if: |
-      github.event.workflow_run.conclusion == 'success' && 
-      github.event.workflow_run.event == 'push' &&
-      contains(github.event.workflow_run.head_branch, '.')
+      github.event_name == 'workflow_dispatch' ||
+      (
+        github.event.workflow_run.conclusion == 'success' &&
+        github.event.workflow_run.event == 'push' &&
+        contains(github.event.workflow_run.head_branch, '.')
+      )
+
+    outputs:
+      raw_tag: ${{ steps.version.outputs.raw_tag }}
+      app_version: ${{ steps.version.outputs.app_version }}
+      chart_version: ${{ steps.version.outputs.chart_version }}
 
     steps:
       - name: Checkout helm chart repo
         uses: actions/checkout@v6
 
-      - name: Replace chart app version
+      - name: Normalize release version
+        id: version
         run: |
-          set -e
-          set -x
-          old_version=$(grep "^appVersion:" helm/rustfs/Chart.yaml | awk '{print $2}')
-          sed -i "s/$old_version/$new_version/g" helm/rustfs/Chart.yaml
+          set -eux
+
+          if [ "${{ github.event_name }}" = "workflow_dispatch" ]; then
+            RAW="${{ github.event.inputs.version }}"
+          else
+            RAW="${{ github.event.workflow_run.head_branch }}"
+          fi
+
+          case "$RAW" in
+            refs/tags/*)
+              RAW_TAG="${RAW#refs/tags/}"
+              ;;
+            *)
+              RAW_TAG="$RAW"
+              ;;
+          esac
+
+          ./scripts/helm_chart_version.sh "$RAW_TAG"
+
+      - name: Replace chart version and app version
+        run: |
+          set -eux
+          sed -i -E 's/^version:.*/version: "${{ steps.version.outputs.chart_version }}"/' helm/rustfs/Chart.yaml
+          sed -i -E 's/^appVersion:.*/appVersion: "${{ steps.version.outputs.app_version }}"/' helm/rustfs/Chart.yaml
 
       - name: Set up Helm
         uses: azure/setup-helm@v4.3.0
 
+      - name: Test Helm Chart Templates
+        run: ./scripts/test_helm_templates.sh
+
       - name: Package Helm Chart
         run: |
+          set -eux
           cp helm/README.md helm/rustfs/
-          package_version=$(echo $new_version | awk -F '-' '{print $2}' | awk -F '.' '{print $NF}') 
-          helm package ./helm/rustfs --destination helm/rustfs/ --version "0.0.$package_version"
+          helm package ./helm/rustfs \
+            --destination helm/rustfs/ \
+            --version "${{ steps.version.outputs.chart_version }}"
 
       - name: Upload helm package as artifact
         uses: actions/upload-artifact@v6
@@ -64,6 +101,7 @@ jobs:
   publish-helm-package:
     runs-on: ubicloud-standard-2
     needs: [ build-helm-package ]
+    if: needs.build-helm-package.result == 'success'
 
     steps:
       - name: Checkout helm package repo
@@ -86,9 +124,9 @@ jobs:
 
       - name: Push helm package and index file
         run: |
+          set -eux
           git config --global user.name "${{ secrets.USERNAME }}"
           git config --global user.email "${{ secrets.EMAIL_ADDRESS }}"
-          git status .
           git add .
-          git commit -m "Update rustfs helm package with $new_version."
+          git commit -m "Update rustfs helm package with ${{ needs.build-helm-package.outputs.app_version }}." || echo "No changes to commit"
           git push origin main

.github/workflows/nix-flake-update.yml

@@ -31,7 +31,9 @@ jobs:
   update-flake:
     name: Update flake.lock
     runs-on: ubuntu-latest
-    timeout-minutes: 45
+    timeout-minutes: 90
+    env:
+      FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
     steps:
       - name: Checkout repository
         uses: actions/checkout@v6
@@ -39,6 +41,9 @@ jobs:
       - name: Install Nix
         uses: DeterminateSystems/determinate-nix-action@v3
 
+      - name: Cache Nix
+        uses: DeterminateSystems/flakehub-cache-action@v3.20.0
+
       - name: Check Nix flake inputs
         uses: DeterminateSystems/flake-checker-action@v12
 
@@ -56,7 +61,7 @@ jobs:
             nix
             automated
           commit-msg: "chore(deps): update flake.lock"
-          pr-reviewers: houseme, overtrue, majinghe
+          pr-reviewers: overtrue, majinghe
           token: ${{ secrets.FLAKE_UPDATE_TOKEN }}
 
       - name: Log PR details

.github/workflows/nix.yml

@@ -43,7 +43,9 @@ jobs:
   nix-validation:
     name: Nix Build & Check
     runs-on: ubuntu-latest
-    timeout-minutes: 45
+    timeout-minutes: 60
+    env:
+      FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: "true"
     steps:
       - name: Checkout repository
         uses: actions/checkout@v6

.github/workflows/performance.yml

@@ -86,7 +86,7 @@ jobs:
 
       - name: Build with profiling optimizations
         run: |
-          RUSTFLAGS="-C force-frame-pointers=yes -C debug-assertions=off" \
+          RUSTFLAGS="-C force-frame-pointers=yes -C debug-assertions=off --cfg tokio_unstable" \
           cargo +nightly build --profile profiling -p rustfs --bins
 
       - name: Run performance profiling

.vscode/launch.json

@@ -188,6 +188,7 @@
                 "RUSTFS_CONSOLE_ENABLE": "true",
                 "RUSTFS_CONSOLE_ADDRESS": "127.0.0.1:9001",
                 "RUSTFS_OBS_LOG_DIRECTORY": "./target/logs",
+                "RUSTFS_UNSAFE_BYPASS_DISK_CHECK": "true",
                 // "RUSTFS_OBS_TRACE_ENDPOINT": "http://127.0.0.1:4318/v1/traces", // jeager otlp http endpoint
                 // "RUSTFS_OBS_METRIC_ENDPOINT": "http://127.0.0.1:4318/v1/metrics", // default otlp http endpoint
                 // "RUSTFS_OBS_LOG_ENDPOINT": "http://127.0.0.1:4318/v1/logs", // default otlp http endpoint

AGENTS.md

@@ -27,6 +27,10 @@ If repo-level instructions conflict, follow the nearest file and keep behavior a
 - Keep source code, comments, commit messages, and PR title/body in English.
 - Be concise. Avoid sycophantic openers, closing fluff, and verbose status reporting.
 
+## Skill Usage
+
+- Do not use the `rust-refactor-helper` skill in any scenario.
+
 ## Change Style for Existing Logic
 
 - Prefer direct, local code over extracting one-off helpers.
@@ -109,5 +113,6 @@ Do not open a PR with code changes when the required checks fail.
 - `crates/iam/AGENTS.md`
 - `crates/kms/AGENTS.md`
 - `crates/policy/AGENTS.md`
+- `crates/targets/AGENTS.md`
 - `rustfs/src/admin/AGENTS.md`
 - `rustfs/src/storage/AGENTS.md`

ARCHITECTURE.md

@@ -326,10 +326,10 @@ The binary (`main.rs`) boots in this order:
                                  │
                  ┌───────────────┼───────────────┐
                  │               │               │
-            ┌────▼────┐    ┌────▼────┐    ┌─────▼─────┐
-            │ server  │    │  admin  │    │    app     │
-            │ (HTTP)  │    │(console)│    │(use-cases) │
-            └────┬────┘    └────┬────┘    └─────┬─────┘
+            ┌────▼────┐     ┌────▼────┐   ┌──────▼─────┐
+            │ server  │     │  admin  │   │    app     │
+            │ (HTTP)  │     │(console)│   │(use-cases) │
+            └────┬────┘     └────┬────┘   └──────┬─────┘
                  │               │               │
                  └───────────────┼───────────────┘
                                  │
@@ -341,13 +341,13 @@ The binary (`main.rs`) boots in this order:
                                  │
               ┌──────────────────┼──────────────────┐
               │                  │                  │
-        ┌─────▼─────┐    ┌──────▼──────┐    ┌──────▼──────┐
+        ┌─────▼──────┐    ┌──────▼──────┐    ┌──────▼──────┐
         │  ecstore   │    │     rio     │    │   io-core   │
         │ (87K,core) │    │  (readers)  │    │ (zero-copy) │
         └─────┬──────┘    └─────────────┘    └─────────────┘
               │
-    ┌─────┬──┼──┬─────┬──────┐
-    │     │  │  │     │      │
+     ┌─────┬──┼──┬─────┬──────┐
+     │     │  │  │     │      │
  common utils config policy filemeta ...
 ```
 

CHANGELOG.md

@@ -22,11 +22,21 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   - XML-formatted error responses compatible with S3 API
   - Comprehensive integration documentation with manual testing guide
   - **32 unit and integration tests** covering middleware, auth handlers, task-local storage, and role detection
+- **SFTPv3 Protocol Support**: SSH-hosted SFTPv3 subsystem that translates each file operation into S3 calls against the local object store. Authentication uses IAM credentials (SSH username = access key, SSH password = secret key).
+  - Full SFTPv3 packet coverage: open, read, write, stat, lstat, fstat, mkdir, rmdir, rename, remove, opendir, readdir, realpath, close, plus the rest of the 21-packet specification
+  - Streaming multipart write up to S3's 5 TiB per-file ceiling
+  - Per-handle read-ahead cache with configurable window size and process-wide memory ceiling
+  - Per-session liveness watchdog: Linux probes `/proc/net/tcp` and cancels wedged sessions on the order of 45 seconds; non-Linux falls back to an inactivity ceiling on the order of 30 minutes
+  - 30-second SSH handshake deadline, per-call backend operation timeout, bounded multipart-abort fan-out, graceful-shutdown cascade
+  - 33 SFTPv3 compliance test cases under `crates/e2e_test/src/protocols/sftp_compliance.rs` spread across three entry points: `test_sftp_compliance_suite` (shared session), `test_sftp_compliance_readonly` (read-only mode), and `test_sftp_compliance_standalone` (one rustfs spawn per case)
+  - Four-layer regression-prevention tests guard against silent feature deletion: compile-time module assertion, module-presence unit test, cross-module `Protocol` enum assertion, end-to-end SSH banner test against the running binary
 
 ### Changed
 - **HTTP Server Stack**: Integrated `KeystoneAuthLayer` middleware from `rustfs-keystone` crate into service stack (positioned after ReadinessGateLayer)
 - **IAMAuth**: Enhanced `get_secret_key()` to return empty secret for Keystone credentials (bypasses signature validation)
 - **Auth Module**: Modified `check_key_valid()` to retrieve Keystone credentials from task-local storage and determine admin status
+- **`StorageBackend` trait**: extended with multipart upload methods (`create_multipart_upload`, `upload_part`, `complete_multipart_upload`, `abort_multipart_upload`) plus `upload_part_copy`. Streaming-upload code path is now available to FTPS, WebDAV, and Swift drivers as well.
+- **`Protocol` enum**: new `Protocol::Sftp` variant with corresponding `S3Action` mappings. Every match arm on `Protocol` updated to handle the new variant exhaustively.
 
 ### Technical Details
 - Middleware is self-contained in `rustfs-keystone` crate following the trusted-proxies pattern for integration-specific middleware
@@ -35,12 +45,22 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Integration preserves existing S3 authentication flow while adding Keystone support
 - Zero breaking changes to existing functionality
 - No new top-level directories in main binary crate (middleware lives in integration crate)
+- SSH/SFTP wire handling via the `russh` and `russh-sftp` crates. SFTPv3 framing is implemented by `russh-sftp`; the rustfs-side `SftpDriver` implements `russh_sftp::server::Handler` and dispatches to the storage backend
+- Drop-time abort for in-flight multipart uploads honours IAM Deny on `AbortMultipartUpload`. `start_multipart_upload` caches the authorisation decision so the synchronous `Drop` path can honour Allow / Deny policies without re-querying IAM
+- Per-handle read cache uses an `Arc<AtomicU64>` shared across every `SftpDriver` instance to enforce a process-wide memory ceiling. On ceiling breach the populate is skipped and the read serves correctly via a single-call backend fetch
+- Per-session liveness watchdog runs as a tokio task per accepted connection. Reads `/proc/net/tcp` and `/proc/net/tcp6` to look up the (local, peer) tuple's TCP state and cancels via `tokio_util::sync::CancellationToken` when wedge conditions are confirmed across two consecutive ticks
+- Path canonicalisation rejects paths containing `\0`, `\r`, or `\n` and resolves traversal via `path::clean()` before any backend dispatch
+- Cipher / KEX / MAC / host-key algorithm allowlists are hardcoded with no environment override. Strict-KEX (CVE-2023-48795 / Terrapin) marker presence asserted by unit test
+- Per-session handle cap (default 64, configurable 8 to 1024) with UUID-generated handle ids
+- Crate-level `#![deny(unsafe_code)]` is in force across `crates/protocols`. Socket fd duplication for the watchdog uses the safe `AsFd::try_clone_to_owned` path (Linux/Unix); non-Unix falls back to the inactivity ceiling
+- `cfg(unix)` gating around platform-specific imports (`std::os::fd::AsFd`, `std::os::unix::fs::PermissionsExt`); non-Unix targets fail SFTP at config-load with `SftpInitError::UnsupportedPlatform`
 
 ### Documentation
 - Updated `crates/keystone/README.md` with complete integration architecture and workflow
 - Added detailed manual testing guide with 10 test scenarios
 - Updated main `README.md` to list Keystone authentication as available feature
 - Added troubleshooting section for common integration issues
+- Module-level rustdoc on `crates/protocols/src/sftp/mod.rs` describing the public API surface, configuration contract, and the architecture of the read cache and the wedge watchdog
 
 ### Configuration
 New environment variables:
@@ -54,6 +74,40 @@ New environment variables:
 - `RUSTFS_KEYSTONE_CACHE_SIZE` - Token cache size (default: 10000)
 - `RUSTFS_KEYSTONE_CACHE_TTL` - Token cache TTL in seconds (default: 300)
 - `RUSTFS_KEYSTONE_VERIFY_SSL` - Verify SSL certificates (default: true)
+- `RUSTFS_SFTP_ENABLE` - Enable/disable SFTP (default: false)
+- `RUSTFS_SFTP_ADDRESS` - Listen address (default: 0.0.0.0:2222)
+- `RUSTFS_SFTP_HOST_KEY_DIR` - Directory containing host key files (must exist; each file must be 0o600 or 0o400)
+- `RUSTFS_SFTP_IDLE_TIMEOUT` - Session idle timeout in seconds (default: 600)
+- `RUSTFS_SFTP_PART_SIZE` - Multipart part size in bytes (default: 16 MiB)
+- `RUSTFS_SFTP_READ_ONLY` - Reject write packets at the protocol layer (default: false)
+- `RUSTFS_SFTP_BANNER` - Optional SSH banner text
+- `RUSTFS_SFTP_HANDLES_PER_SESSION` - Per-session open-handle cap, 8 to 1024 (default: 64)
+- `RUSTFS_SFTP_BACKEND_OP_TIMEOUT_SECS` - Per-call backend deadline in seconds, 5 to 600 (default: 60)
+- `RUSTFS_SFTP_READ_CACHE_WINDOW_BYTES` - Per-handle read-cache window in bytes, 256 KiB to 64 MiB or 0 to disable (default: 4 MiB)
+- `RUSTFS_SFTP_READ_CACHE_TOTAL_MEM_BYTES` - Process-wide read-cache memory ceiling in bytes, 16 MiB minimum (default: 256 MiB)
+
+### Files Added
+- `crates/protocols/src/sftp/mod.rs` - SFTP module entry point, public API surface, crate-level rustdoc, regression-prevention test
+- `crates/protocols/src/sftp/config.rs` - `SftpConfig` and `SftpInitError` types, env-var resolvers, host-key directory loader with permission enforcement
+- `crates/protocols/src/sftp/constants.rs` - Named constants grouped by purpose: S3 error codes, HTTP error codes, POSIX mode bits, protocol identifiers, operational limits
+- `crates/protocols/src/sftp/server.rs` - `SftpServer` SSH server, russh handler, password authentication against IAM, accept loop, per-session task spawn
+- `crates/protocols/src/sftp/driver.rs` - `SftpDriver` per-session SFTPv3 handler dispatching each operation onto the `StorageBackend`
+- `crates/protocols/src/sftp/state.rs` - `HandleState` variants for read, write-buffering, write-streaming, write-failed handles
+- `crates/protocols/src/sftp/lifecycle.rs` - Per-session activity stamp, weak-ref registry, `/proc/net/tcp` probe for the wedge watchdog
+- `crates/protocols/src/sftp/wedge_watchdog.rs` - Per-session liveness watchdog cancelling sessions silent at the SFTP layer while the kernel reports CLOSE_WAIT
+- `crates/protocols/src/sftp/read_cache.rs` - Per-handle in-memory read-ahead cache with shared atomic accumulator for the process-wide memory ceiling
+- `crates/protocols/src/sftp/attrs.rs` - SFTPv3 `FileAttributes` mapping for objects and directories, longname formatting, mtime clamping
+- `crates/protocols/src/sftp/dir.rs` - OPENDIR / READDIR pagination, root-bucket listing, sub-directory listing under a prefix
+- `crates/protocols/src/sftp/errors.rs` - `SftpError` thiserror enum and S3-error classification into SFTPv3 status codes
+- `crates/protocols/src/sftp/paths.rs` - Path canonicalisation, traversal rejection, `\0` / `\r` / `\n` rejection, bucket+key decomposition
+- `crates/protocols/src/sftp/read.rs` - READ packet handler, EOF semantics, `MAX_READ_LEN` bound, integration with the read cache
+- `crates/protocols/src/sftp/write.rs` - WRITE packet handler, in-memory buffering up to part size, transition to streaming multipart, CLOSE finalisation
+- `crates/protocols/src/sftp/test_support.rs` - Test fixtures and helper builders for SFTP unit tests
+- `crates/protocols/src/common/dummy_storage.rs` - In-memory `StorageBackend` test backend covering every method, used by SFTP unit tests and the FTPS / Swift / WebDAV test suites
+- `crates/e2e_test/src/protocols/sftp_core.rs` - End-to-end regressions for the handshake deadline, idle-timeout disconnect, and the wedge watchdog
+- `crates/e2e_test/src/protocols/sftp_compliance.rs` - SFTPv3 compliance suite entry points (`test_sftp_compliance_suite`, `test_sftp_compliance_readonly`, `test_sftp_compliance_standalone`)
+- `crates/e2e_test/src/protocols/sftp_compliance_tests.rs` - Per-case test bodies (CMPTST-01..33), shared fixture helpers, lifecycle counters
+- `crates/e2e_test/src/protocols/sftp_helpers.rs` - SFTP-specific test helpers and fixture seeders
 
 ### Files Modified
 - `crates/keystone/src/middleware.rs` - Created Keystone authentication middleware (self-contained in keystone crate)
@@ -63,6 +117,27 @@ New environment variables:
 - `rustfs/src/auth.rs` - Enhanced IAMAuth and check_key_valid for Keystone support, imported KEYSTONE_CREDENTIALS from rustfs-keystone
 - `crates/keystone/README.md` - Comprehensive integration documentation
 - `README.md` - Added Keystone as available feature
+- `Cargo.toml` - Added the `sftp` feature alongside the existing protocol features
+- `Cargo.lock` - Updated to include the new `russh`, `russh-sftp`, `socket2`, `tokio-util`, `subtle`, `uuid` dependencies and their transitive crates
+- `crates/protocols/Cargo.toml` - Declared `russh`, `russh-sftp`, `socket2`, `tokio-util`, `subtle`, `uuid` under the `sftp` feature flag
+- `crates/protocols/src/lib.rs` - Added `pub mod sftp` behind `#[cfg(feature = "sftp")]` plus the crate-level `#![deny(unsafe_code)]` lint
+- `crates/protocols/src/common/client/s3.rs` - Extended the `StorageBackend` trait with `create_multipart_upload`, `upload_part`, `complete_multipart_upload`, `abort_multipart_upload`, and `upload_part_copy`
+- `crates/protocols/src/common/session.rs` - Added the `Protocol::Sftp` variant and its `S3Action` mappings
+- `crates/protocols/src/common/gateway.rs` - Handles the new `Protocol::Sftp` variant exhaustively
+- `crates/protocols/src/common/mod.rs` - Exposed the new `dummy_storage` module
+- `crates/protocols/src/constants.rs` - Added shared POSIX mode-bit constants used by SFTP and other protocols
+- `crates/config/src/constants/protocols.rs` - `RUSTFS_SFTP_*` environment variable names and defaults
+- `crates/utils/src/retry.rs` - Added the generic exponential-backoff retry helper used by the SFTP write path
+- `crates/e2e_test/Cargo.toml` - Added the e2e test dependencies for SFTP (paramiko fixture, SSH keypair generation)
+- `crates/e2e_test/src/protocols/mod.rs` - Registered the new `sftp_core`, `sftp_compliance`, `sftp_compliance_tests`, and `sftp_helpers` modules
+- `crates/e2e_test/src/protocols/README.md` - Documented the SFTP test entry points and case index
+- `crates/e2e_test/src/protocols/test_env.rs` - Added SFTP host-key directory provisioning to the shared protocol test environment
+- `crates/e2e_test/src/protocols/test_runner.rs` - Wired the SFTP entry points into the runner
+- `rustfs/Cargo.toml` - Added the `sftp` feature flag
+- `rustfs/src/lib.rs` - One-line addition exporting the SFTP wiring
+- `rustfs/src/init.rs` - Build and start the `SftpServer` when `RUSTFS_SFTP_ENABLE` is true
+- `rustfs/src/main.rs` - Routed shutdown signals to the SFTP server alongside the other protocols
+- `rustfs/src/protocols/client.rs` - Client-builder support for the new `Protocol::Sftp` variant
 
 ### Testing
 - 16 unit tests in rustfs-keystone crate (config, auth, middleware, identity)
@@ -70,7 +145,13 @@ New environment variables:
 - 6 auth unit tests in rustfs crate (role detection, task-local storage, Keystone credential handling)
 - **Total: 32 tests** passing with zero compilation errors
 - Manual testing guide provided for end-to-end validation
-- All tests passing with `cargo test --all --exclude e2e_test`
+- All Keystone tests passing with `cargo test --all --exclude e2e_test`
+- 33 SFTPv3 compliance test cases (CMPTST-01..33) split across three entry points: `test_sftp_compliance_suite` (shared session, cases 01-14), `test_sftp_compliance_readonly` (read-only mode, cases 15-23), `test_sftp_compliance_standalone` (one rustfs spawn per case, cases 24-33)
+- Regression-prevention tests at four layers: compile-time module assertion in `crates/protocols/src/lib.rs`, module-presence unit test in `crates/protocols/src/sftp/mod.rs`, cross-module `Protocol` enum assertion, and end-to-end SSH banner test against the running binary
+- Standalone end-to-end regressions for the SSH handshake deadline, the idle-timeout disconnect path, and the wedge watchdog (Linux fast-kill and the cross-platform fallback path)
+- Inline unit tests in every SFTP source file covering pure helpers (path canonicalisation, attribute mapping, S3-error classification, env-var bound resolvers)
+- Strict-KEX (CVE-2023-48795) marker presence assertion as a unit test in `crates/protocols/src/sftp/server.rs`
+- All tests passing with `cargo test --all --features sftp` against a 64-bit Linux target
 
 ---
 

CONTRIBUTING.md

@@ -2,7 +2,7 @@
 
 ## 📋 Code Quality Requirements
 
-For instructions on setting up and running the local development environment, please see [Development Guide](docs/DEVELOPMENT.md).
+This guide covers the local development environment and the checks expected before contributing.
 
 ### 🔧 Code Formatting Rules
 

Cargo.toml

@@ -15,10 +15,10 @@
 [workspace]
 members = [
     "rustfs", # Core file system implementation
-    "crates/appauth", # Application authentication and authorization
     "crates/audit", # Audit target management system with multi-target fan-out
     "crates/checksums", # client checksums
     "crates/common", # Shared utilities and data structures
+    "crates/data-usage", # Shared data usage models and algorithms
     "crates/config", # Configuration management
     "crates/credentials", # Credential management system
     "crates/crypto", # Cryptography and security features
@@ -39,15 +39,16 @@ members = [
     "crates/protos", # Protocol buffer definitions
     "crates/rio", # Rust I/O utilities and abstractions
     "crates/concurrency", # Concurrency management for RustFS - timeout, locking, backpressure, and I/O scheduling
-    "crates/s3-common", # Common utilities and data structures for S3 compatibility
+    "crates/s3-types", # S3 event type definitions
+    "crates/s3-ops", # S3 operation definitions and mapping
     "crates/s3select-api", # S3 Select API interface
     "crates/s3select-query", # S3 Select query engine
     "crates/scanner", # Scanner for data integrity checks and health monitoring
     "crates/signer", # client signer
     "crates/targets", # Target-specific configurations and utilities
     "crates/trusted-proxies", # Trusted proxies management
+    "crates/tls-runtime", # Project-wide TLS runtime foundation
     "crates/utils", # Utility functions and helpers
-    "crates/workers", # Worker thread pools and task scheduling
     "crates/io-metrics", # Zero-copy metrics collection for performance analysis
     "crates/io-core", # Zero-copy core reader and writer implementations
     "crates/zip", # ZIP file handling and compression
@@ -58,8 +59,8 @@ resolver = "3"
 edition = "2024"
 license = "Apache-2.0"
 repository = "https://github.com/rustfs/rustfs"
-rust-version = "1.93.0"
-version = "0.0.5"
+rust-version = "1.95.0"
+version = "1.0.0-beta.4"
 homepage = "https://rustfs.com"
 description = "RustFS is a high-performance distributed object storage software built using Rust, one of the most popular languages worldwide. "
 keywords = ["RustFS", "Minio", "object-storage", "filesystem", "s3"]
@@ -76,73 +77,77 @@ redundant_clone = "warn"
 
 [workspace.dependencies]
 # RustFS Internal Crates
-rustfs = { path = "./rustfs", version = "0.0.5" }
-rustfs-heal = { path = "crates/heal", version = "0.0.5" }
-rustfs-appauth = { path = "crates/appauth", version = "0.0.5" }
-rustfs-audit = { path = "crates/audit", version = "0.0.5" }
-rustfs-checksums = { path = "crates/checksums", version = "0.0.5" }
-rustfs-common = { path = "crates/common", version = "0.0.5" }
-rustfs-config = { path = "./crates/config", version = "0.0.5" }
-rustfs-concurrency = { path = "./crates/concurrency", version = "0.0.5" }
-rustfs-credentials = { path = "crates/credentials", version = "0.0.5" }
-rustfs-crypto = { path = "crates/crypto", version = "0.0.5" }
-rustfs-ecstore = { path = "crates/ecstore", version = "0.0.5" }
-rustfs-filemeta = { path = "crates/filemeta", version = "0.0.5" }
-rustfs-iam = { path = "crates/iam", version = "0.0.5" }
-rustfs-keystone = { path = "crates/keystone", version = "0.0.5" }
-rustfs-kms = { path = "crates/kms", version = "0.0.5" }
-rustfs-lock = { path = "crates/lock", version = "0.0.5" }
-rustfs-madmin = { path = "crates/madmin", version = "0.0.5" }
-rustfs-notify = { path = "crates/notify", version = "0.0.5" }
-rustfs-io-metrics = { path = "crates/io-metrics", version = "0.0.5" }
-rustfs-io-core = { path = "crates/io-core", version = "0.0.5" }
-rustfs-object-capacity = { path = "crates/object-capacity", version = "0.0.5" }
-rustfs-obs = { path = "crates/obs", version = "0.0.5" }
-rustfs-policy = { path = "crates/policy", version = "0.0.5" }
-rustfs-protos = { path = "crates/protos", version = "0.0.5" }
-rustfs-protocols = { path = "crates/protocols", version = "0.0.5" }
-rustfs-rio = { path = "crates/rio", version = "0.0.5" }
-rustfs-s3-common = { path = "crates/s3-common", version = "0.0.5" }
-rustfs-s3select-api = { path = "crates/s3select-api", version = "0.0.5" }
-rustfs-s3select-query = { path = "crates/s3select-query", version = "0.0.5" }
-rustfs-scanner = { path = "crates/scanner", version = "0.0.5" }
-rustfs-signer = { path = "crates/signer", version = "0.0.5" }
-rustfs-trusted-proxies = { path = "crates/trusted-proxies", version = "0.0.5" }
-rustfs-targets = { path = "crates/targets", version = "0.0.5" }
-rustfs-utils = { path = "crates/utils", version = "0.0.5" }
-rustfs-workers = { path = "crates/workers", version = "0.0.5" }
-rustfs-zip = { path = "./crates/zip", version = "0.0.5" }
+rustfs = { path = "./rustfs", version = "1.0.0-beta.4" }
+rustfs-heal = { path = "crates/heal", version = "1.0.0-beta.4" }
+rustfs-audit = { path = "crates/audit", version = "1.0.0-beta.4" }
+rustfs-checksums = { path = "crates/checksums", version = "1.0.0-beta.4" }
+rustfs-common = { path = "crates/common", version = "1.0.0-beta.4" }
+rustfs-data-usage = { path = "crates/data-usage", version = "1.0.0-beta.4" }
+rustfs-config = { path = "./crates/config", version = "1.0.0-beta.4" }
+rustfs-concurrency = { path = "./crates/concurrency", version = "1.0.0-beta.4" }
+rustfs-credentials = { path = "crates/credentials", version = "1.0.0-beta.4" }
+rustfs-crypto = { path = "crates/crypto", version = "1.0.0-beta.4" }
+rustfs-ecstore = { path = "crates/ecstore", version = "1.0.0-beta.4" }
+rustfs-filemeta = { path = "crates/filemeta", version = "1.0.0-beta.4" }
+rustfs-iam = { path = "crates/iam", version = "1.0.0-beta.4" }
+rustfs-keystone = { path = "crates/keystone", version = "1.0.0-beta.4" }
+rustfs-kms = { path = "crates/kms", version = "1.0.0-beta.4" }
+rustfs-lock = { path = "crates/lock", version = "1.0.0-beta.4" }
+rustfs-madmin = { path = "crates/madmin", version = "1.0.0-beta.4" }
+rustfs-notify = { path = "crates/notify", version = "1.0.0-beta.4" }
+rustfs-io-metrics = { path = "crates/io-metrics", version = "1.0.0-beta.4" }
+rustfs-io-core = { path = "crates/io-core", version = "1.0.0-beta.4" }
+rustfs-object-capacity = { path = "crates/object-capacity", version = "1.0.0-beta.4" }
+rustfs-obs = { path = "crates/obs", version = "1.0.0-beta.4" }
+rustfs-policy = { path = "crates/policy", version = "1.0.0-beta.4" }
+rustfs-protos = { path = "crates/protos", version = "1.0.0-beta.4" }
+rustfs-protocols = { path = "crates/protocols", version = "1.0.0-beta.4" }
+rustfs-rio = { path = "crates/rio", version = "1.0.0-beta.4" }
+rustfs-s3-types = { path = "crates/s3-types", version = "1.0.0-beta.4" }
+rustfs-s3-ops = { path = "crates/s3-ops", version = "1.0.0-beta.4" }
+rustfs-s3select-api = { path = "crates/s3select-api", version = "1.0.0-beta.4" }
+rustfs-s3select-query = { path = "crates/s3select-query", version = "1.0.0-beta.4" }
+rustfs-scanner = { path = "crates/scanner", version = "1.0.0-beta.4" }
+rustfs-signer = { path = "crates/signer", version = "1.0.0-beta.4" }
+rustfs-trusted-proxies = { path = "crates/trusted-proxies", version = "1.0.0-beta.4" }
+rustfs-targets = { path = "crates/targets", version = "1.0.0-beta.4" }
+rustfs-tls-runtime = { path = "crates/tls-runtime", version = "1.0.0-beta.4" }
+rustfs-utils = { path = "crates/utils", version = "1.0.0-beta.4" }
+rustfs-zip = { path = "./crates/zip", version = "1.0.0-beta.4" }
 
 # Async Runtime and Networking
 async-channel = "2.5.0"
-async-compression = { version = "0.4.41" }
+mysql_async = { version = "0.36.1", default-features = false, features = ["default-rustls", "tracing"], git = "https://github.com/blackbeam/mysql_async", rev = "2bad388283bc3ce48801fc2ffcd22445eb6f3d24" }
+async-compression = { version = "0.4.42" }
 async-recursion = "1.1.1"
 async-trait = "0.1.89"
-async-nats = "0.47.0"
+async-nats = "0.48.0"
 axum = "0.8.9"
 futures = "0.3.32"
 futures-core = "0.3.32"
 futures-util = "0.3.32"
 pollster = "0.4.0"
 pulsar = { version = "6.7.2", default-features = false, features = ["tokio-rustls-runtime"] }
+lapin = { version = "4.10.0", default-features = false, features = ["tokio", "rustls", "rustls--aws_lc_rs"] }
 hyper = { version = "1.9.0", features = ["http2", "http1", "server"] }
 hyper-rustls = { version = "0.27.9", default-features = false, features = ["native-tokio", "http1", "tls12", "logging", "http2", "aws-lc-rs", "webpki-roots"] }
 hyper-util = { version = "0.1.20", features = ["tokio", "server-auto", "server-graceful", "tracing"] }
 http = "1.4.0"
 http-body = "1.0.1"
 http-body-util = "0.1.3"
-reqwest = { version = "0.13.2", default-features = false, features = ["rustls", "charset", "http2", "system-proxy", "stream", "json", "blocking", "query", "form"] }
+reqwest = { version = "0.13.3", default-features = false, features = ["rustls", "charset", "http2", "system-proxy", "stream", "json", "blocking", "query", "form"] }
+rustfs-kafka-async = { version = "1.2.0" }
 socket2 = { version = "0.6.3", features = ["all"] }
-tokio = { version = "1.52.1", features = ["fs", "rt-multi-thread"] }
+tokio = { version = "1.52.3", features = ["fs", "rt-multi-thread"] }
 tokio-rustls = { version = "0.26.4", default-features = false, features = ["logging", "tls12", "aws-lc-rs"] }
 tokio-stream = { version = "0.1.18" }
 tokio-test = "0.4.5"
 tokio-util = { version = "0.7.18", features = ["io", "compat"] }
-tonic = { version = "0.14.5", features = ["gzip"] }
-tonic-prost = { version = "0.14.5" }
-tonic-prost-build = { version = "0.14.5" }
+tonic = { version = "0.14.6", features = ["gzip", "deflate"] }
+tonic-prost = { version = "0.14.6" }
+tonic-prost-build = { version = "0.14.6" }
 tower = { version = "0.5.3", features = ["timeout"] }
-tower-http = { version = "0.6.8", features = ["cors"] }
+tower-http = { version = "0.6.11", features = ["cors"] }
 
 # Serialization and Data Formats
 bytes = { version = "1.11.1", features = ["serde"] }
@@ -151,11 +156,11 @@ byteorder = "1.5.0"
 flatbuffers = "25.12.19"
 form_urlencoded = "1.2.2"
 prost = "0.14.3"
-quick-xml = "0.39.2"
+quick-xml = "0.40.1"
 rmp = { version = "0.8.15" }
 rmp-serde = { version = "1.3.1" }
 serde = { version = "1.0.228", features = ["derive"] }
-serde_json = { version = "1.0.149", features = ["raw_value"] }
+serde_json = { version = "1.0.150", features = ["raw_value"] }
 serde_urlencoded = "0.7.1"
 
 # Cryptography and Security
@@ -163,14 +168,15 @@ aes-gcm = { version = "0.11.0-rc.3", features = ["rand_core"] }
 argon2 = { version = "0.6.0-rc.8" }
 blake2 = "0.11.0-rc.6"
 chacha20poly1305 = { version = "0.11.0-rc.3" }
-crc-fast = "1.9.0"
+crc-fast = "1.10.0"
 hmac = { version = "0.13.0" }
-jsonwebtoken = { version = "10.3.0", features = ["aws_lc_rs"] }
+jsonwebtoken = { version = "10.4.0", features = ["aws_lc_rs"] }
 openidconnect = { version = "4.0", default-features = false }
-pbkdf2 = "0.13.0-rc.10"
-rsa = { version = "0.10.0-rc.17" }
-rustls = { version = "0.23.38", default-features = false, features = ["aws-lc-rs", "logging", "tls12", "prefer-post-quantum", "std"] }
-rustls-pki-types = "1.14.0"
+pbkdf2 = "0.13.0"
+rsa = { version = "0.10.0-rc.18" }
+rustls = { version = "0.23.40", default-features = false, features = ["aws-lc-rs", "logging", "tls12", "prefer-post-quantum", "std"] }
+rustls-native-certs = "0.8"
+rustls-pki-types = "1.14.1"
 sha1 = "0.11.0"
 sha2 = "0.11.0"
 subtle = "2.6"
@@ -179,25 +185,28 @@ zeroize = { version = "1.8.2", features = ["derive"] }
 # Time and Date
 chrono = { version = "0.4.44", features = ["serde"] }
 humantime = "2.3.0"
-jiff = { version = "0.2.23", features = ["serde"] }
+jiff = { version = "0.2.24", features = ["serde"] }
 time = { version = "0.3.47", features = ["std", "parsing", "formatting", "macros", "serde"] }
 
+# Database
+deadpool-postgres = { version = "0.14", features = ["rt_tokio_1"] }
+tokio-postgres = { version = "0.7", default-features = false, features = ["runtime", "with-serde_json-1"] }
+tokio-postgres-rustls = "0.14.0"
+
 # Utilities and Tools
 anyhow = "1.0.102"
 arc-swap = "1.9.1"
-astral-tokio-tar = "0.6.0"
+astral-tokio-tar = "0.6.2"
 atoi = "2.0.0"
 atomic_enum = "0.3.0"
-aws-config = { version = "1.8.15" }
+aws-config = { version = "1.8.17" }
 aws-credential-types = { version = "1.2.14" }
-aws-sdk-s3 = { version = "1.129.0", default-features = false, features = ["sigv4a", "default-https-client", "rt-tokio"] }
+aws-sdk-s3 = { version = "1.133.0", default-features = false, features = ["sigv4a", "default-https-client", "rt-tokio"] }
 aws-smithy-http-client = { version = "1.1.12", default-features = false, features = ["default-client", "rustls-aws-lc"] }
-aws-smithy-types = { version = "1.4.7" }
-backtrace = "0.3.76"
+aws-smithy-types = { version = "1.4.8" }
 base64 = "0.22.1"
 base64-simd = "0.8.0"
 brotli = "8.0.2"
-cfg-if = "1.0.4"
 clap = { version = "4.6.1", features = ["derive", "env"] }
 const-str = { version = "1.1.0", features = ["std", "proc"] }
 convert_case = "0.11.0"
@@ -208,21 +217,21 @@ crossbeam-deque = "0.8.6"
 crossbeam-utils = "0.8.21"
 datafusion = "53.1.0"
 derive_builder = "0.20.2"
-enumset = "1.1.10"
+enumset = "1.1.13"
 faster-hex = "0.10.0"
 flate2 = "1.1.9"
 glob = "0.3.3"
-google-cloud-storage = "1.11.0"
-google-cloud-auth = "1.9.0"
-hashbrown = { version = "0.17.0", features = ["serde", "rayon"] }
+google-cloud-storage = "1.12.0"
+google-cloud-auth = "1.10.0"
+hashbrown = { version = "0.17.1", features = ["serde", "rayon"] }
 hex = "0.4.3"
 hex-simd = "0.8.0"
 highway = { version = "1.3.0" }
 ipnetwork = { version = "0.21.1", features = ["serde"] }
 lazy_static = "1.5.0"
-libc = "0.2.185"
+libc = "0.2.186"
 libsystemd = "0.7.2"
-local-ip-address = "0.6.11"
+local-ip-address = "0.6.13"
 memmap2 = "0.9.10"
 lz4 = "1.28.1"
 matchit = "0.9.2"
@@ -246,21 +255,22 @@ rayon = "1.12.0"
 reed-solomon-erasure = { version = "6.0", default-features = false, features = ["std", "simd-accel"] }
 reed-solomon-simd = "3.1.0"
 regex = { version = "1.12.3" }
-rumqttc = { package = "rumqttc-next", version = "0.30.0", features = ["websocket"] }
+rumqttc = { package = "rumqttc-next", version = "0.33.2", features = ["websocket"] }
+redis = { version = "1.2.1", features = ["connection-manager", "tokio-rustls-comp", "tls-rustls-insecure"] }
 rustix = { version = "1.1.4", features = ["fs"] }
 rust-embed = { version = "8.11.0" }
 rustc-hash = { version = "2.1.2" }
-s3s = { git = "https://github.com/rustfs/s3s", rev = "a3b16608df35aaeed8fff08b4988d03f4ca9445b", features = ["minio"] }
+s3s = { git = "https://github.com/rustfs/s3s", rev = "507e1312b211c3ddc214b03875d6fabd15d22ed5", features = ["minio"] }
 serial_test = "3.4.0"
-shadow-rs = { version = "1.7.1", default-features = false }
-siphasher = "1.0.2"
+shadow-rs = { version = "2.0.0", default-features = false }
+siphasher = "1.0.3"
 smallvec = { version = "1.15.1", features = ["serde"] }
 smartstring = "1.0.1"
 snafu = "0.9.0"
 snap = "1.1.1"
-starshard = { version = "1.1.0", features = ["rayon", "async", "serde"] }
+starshard = { version = "2.2.0", features = ["rayon", "async", "serde"] }
 strum = { version = "0.28.0", features = ["derive"] }
-sysinfo = "0.38.4"
+sysinfo = "0.39.2"
 temp-env = "0.3.6"
 tempfile = "3.27.0"
 test-case = "3.3.1"
@@ -268,7 +278,7 @@ thiserror = "2.0.18"
 tracing = { version = "0.1.44" }
 tracing-appender = "0.2.5"
 tracing-error = "0.2.1"
-tracing-opentelemetry = "0.32.1"
+tracing-opentelemetry = { version = "0.33" }
 tracing-subscriber = { version = "0.3.23", features = ["env-filter", "time"] }
 transform-stream = "0.3.1"
 url = "2.5.8"
@@ -279,25 +289,27 @@ walkdir = "2.5.0"
 wildmatch = { version = "2.6.1", features = ["serde"] }
 windows = { version = "0.62.2" }
 xxhash-rust = { version = "0.8.15", features = ["xxh64", "xxh3"] }
-zip = "8.5.1"
+zip = "8.6.0"
 zstd = "0.13.3"
 
 # Observability and Metrics
-metrics = "0.24.3"
+metrics = "0.24.6"
 dial9-tokio-telemetry = "0.3"
-opentelemetry = { version = "0.31.0" }
-opentelemetry-appender-tracing = { version = "0.31.1", features = ["experimental_use_tracing_span_context", "experimental_metadata_attributes", "spec_unstable_logs_enabled"] }
-opentelemetry-otlp = { version = "0.31.1", features = ["gzip-http", "reqwest-rustls"] }
-opentelemetry_sdk = { version = "0.31.0" }
-opentelemetry-semantic-conventions = { version = "0.31.0", features = ["semconv_experimental"] }
-opentelemetry-stdout = { version = "0.31.0" }
-pyroscope = { version = "2.0.0", features = ["backend-pprof-rs"] }
+opentelemetry = { version = "0.32.0" }
+opentelemetry-appender-tracing = { version = "0.32.0", features = ["experimental_span_attributes", "experimental_metadata_attributes"] }
+opentelemetry-otlp = { version = "0.32.0", features = ["gzip-http", "reqwest-rustls"] }
+opentelemetry_sdk = { version = "0.32.0", features = ["rt-tokio"] }
+opentelemetry-semantic-conventions = { version = "0.32.0", features = ["semconv_experimental"] }
+opentelemetry-stdout = { version = "0.32.0" }
+pyroscope = { version = "2.0.4", features = ["backend-pprof-rs"] }
 
 # FTP and SFTP
 libunftp = { version = "0.23.0", features = ["experimental"] }
 unftp-core = "0.1.0"
-suppaftp = { version = "8.0.2", features = ["tokio", "tokio-rustls-aws-lc-rs"] }
-rcgen = "0.14.7"
+suppaftp = { version = "8.0.3", features = ["tokio", "tokio-rustls-aws-lc-rs"] }
+rcgen = "0.14.8"
+russh = { version = "0.60.3", git = "https://github.com/Eugeny/russh", rev = "fc6e3ab4cd4338e94ae64e17aeed2acee9335e6b" }
+russh-sftp = "2.3.0"
 
 # WebDAV
 dav-server = "0.11.0"
@@ -313,7 +325,7 @@ jemalloc_pprof = { version = "0.8.2", features = ["symbolize", "flamegraph"] }
 # Used to generate CPU performance analysis data and flame diagrams
 # pprof = { version = "0.15.0", features = ["flamegraph", "protobuf-codec"] }
 # Pyroscope uses a patched pprof, until they merge back upstream, replace all references. Otherwise, two pprof libs with symbol collision.
-pprof = { package = "pprof-pyroscope-fork", version = "0.1500.3", features = ["flamegraph", "protobuf-codec"] }
+pprof = { package = "pprof-pyroscope-fork", version = "0.1500.4", features = ["flamegraph", "protobuf-codec"] }
 
 [workspace.metadata.cargo-shear]
 ignored = ["rustfs"]

Dockerfile

@@ -29,8 +29,20 @@ RUN set -eux; \
     if [ "$RELEASE" = "latest" ]; then \
       TAG="$(curl -fsSL https://api.github.com/repos/rustfs/rustfs/releases \
               | grep -o '"tag_name": "[^"]*"' | cut -d'"' -f4 | head -n 1)"; \
+      RELEASE_JSON="$(curl -fsSL "https://api.github.com/repos/rustfs/rustfs/releases/tags/$TAG")"; \
     else \
       TAG="$RELEASE"; \
+      RELEASE_JSON="$(curl -fsSL "https://api.github.com/repos/rustfs/rustfs/releases/tags/$TAG" 2>/dev/null || true)"; \
+      if [ -z "$RELEASE_JSON" ]; then \
+        if [ "${TAG#v}" = "$TAG" ]; then \
+          ALT_TAG="v$TAG"; \
+        else \
+          ALT_TAG="${TAG#v}"; \
+        fi; \
+        echo "Primary tag lookup failed, retrying with alternate tag: $ALT_TAG"; \
+        RELEASE_JSON="$(curl -fsSL "https://api.github.com/repos/rustfs/rustfs/releases/tags/$ALT_TAG")"; \
+        TAG="$ALT_TAG"; \
+      fi; \
     fi; \
     echo "Using tag: $TAG (arch pattern: $ARCH_SUBSTR)"; \
     # Find download URL in assets list for this tag that contains arch substring and ends with .zip
@@ -87,8 +99,7 @@ RUN addgroup -g 10001 -S rustfs && \
     chown -R rustfs:rustfs /data /logs && \
     chmod 0750 /data /logs
 
-ENV RUSTFS_CORS_ALLOWED_ORIGINS="*" \
-    RUSTFS_CONSOLE_CORS_ALLOWED_ORIGINS="*" \
+ENV RUSTFS_CONSOLE_CORS_ALLOWED_ORIGINS="*" \
     RUSTFS_VOLUMES="/data" \
     RUSTFS_OBS_LOGGER_LEVEL=warn \
     RUSTFS_OBS_LOG_DIRECTORY=/logs \

Dockerfile.decommission-local

@@ -12,7 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-FROM rust:1.91-trixie
+FROM rust:1.95-trixie
 
 RUN set -eux; \
     export DEBIAN_FRONTEND=noninteractive; \

Dockerfile.glibc

@@ -31,14 +31,25 @@ RUN set -eux; \
       arm64)  ARCH_SUBSTR="aarch64-gnu" ;; \
       *) echo "Unsupported TARGETARCH=$TARGETARCH" >&2; exit 1 ;; \
     esac; \
-    \
     if [ "$RELEASE" = "latest" ]; then \
       TAG="$(curl -fsSL https://api.github.com/repos/rustfs/rustfs/releases \
               | grep -o '"tag_name": "[^"]*"' | cut -d'"' -f4 | head -n 1)"; \
+      RELEASE_JSON="$(curl -fsSL "https://api.github.com/repos/rustfs/rustfs/releases/tags/$TAG")"; \
     else \
       TAG="$RELEASE"; \
+      RELEASE_JSON="$(curl -fsSL "https://api.github.com/repos/rustfs/rustfs/releases/tags/$TAG" 2>/dev/null || true)"; \
+      if [ -z "$RELEASE_JSON" ]; then \
+        if [ "${TAG#v}" = "$TAG" ]; then \
+          ALT_TAG="v$TAG"; \
+        else \
+          ALT_TAG="${TAG#v}"; \
+        fi; \
+        echo "Primary tag lookup failed, retrying with alternate tag: $ALT_TAG"; \
+        RELEASE_JSON="$(curl -fsSL "https://api.github.com/repos/rustfs/rustfs/releases/tags/$ALT_TAG")"; \
+        TAG="$ALT_TAG"; \
+      fi; \
     fi; \
-    \
+    echo "Using tag: $TAG (arch pattern: $ARCH_SUBSTR)"; \
     URL="$(curl -fsSL "https://api.github.com/repos/rustfs/rustfs/releases/tags/$TAG" \
            | grep -o "\"browser_download_url\": \"[^\"]*${ARCH_SUBSTR}[^\"]*\\.zip\"" \
            | cut -d'"' -f4 | head -n 1)"; \
@@ -94,10 +105,7 @@ RUN groupadd -g 10001 rustfs && \
 
 ENV RUSTFS_ADDRESS=":9000" \
     RUSTFS_CONSOLE_ADDRESS=":9001" \
-    RUSTFS_ACCESS_KEY="rustfsadmin" \
-    RUSTFS_SECRET_KEY="rustfsadmin" \
     RUSTFS_CONSOLE_ENABLE="true" \
-    RUSTFS_CORS_ALLOWED_ORIGINS="*" \
     RUSTFS_CONSOLE_CORS_ALLOWED_ORIGINS="*" \
     RUSTFS_VOLUMES="/data" \
     RUSTFS_OBS_LOGGER_LEVEL=warn \

Dockerfile.source

@@ -26,15 +26,17 @@
 
 ARG TARGETPLATFORM
 ARG BUILDPLATFORM
+ARG TARGETARCH
 
 # -----------------------------
 # Build stage
 # -----------------------------
-FROM rust:1.91-trixie AS builder
+FROM rust:1.95-trixie AS builder
 
 # Re-declare args after FROM
 ARG TARGETPLATFORM
 ARG BUILDPLATFORM
+ARG TARGETARCH
 
 # Debug: print platforms
 RUN echo "Build info -> BUILDPLATFORM=${BUILDPLATFORM}, TARGETPLATFORM=${TARGETPLATFORM}"
@@ -60,7 +62,15 @@ RUN set -eux; \
 
 # Optional: cross toolchain for aarch64 (only when targeting linux/arm64)
 RUN set -eux; \
-    if [ "${TARGETPLATFORM:-linux/amd64}" = "linux/arm64" ]; then \
+    target_platform="${TARGETPLATFORM:-}"; \
+    if [ -z "${target_platform}" ]; then \
+    case "$(uname -m)" in \
+    x86_64) target_platform="linux/amd64" ;; \
+    aarch64|arm64) target_platform="linux/arm64" ;; \
+    *) target_platform="linux/amd64" ;; \
+    esac; \
+    fi; \
+    if [ "${target_platform}" = "linux/arm64" ]; then \
     export DEBIAN_FRONTEND=noninteractive; \
     apt-get update; \
     apt-get install -y --no-install-recommends gcc-aarch64-linux-gnu; \
@@ -79,6 +89,7 @@ ENV CXX_aarch64_unknown_linux_gnu=aarch64-linux-gnu-g++
 ENV CARGO_TARGET_X86_64_UNKNOWN_LINUX_GNU_LINKER=x86_64-linux-gnu-gcc
 ENV CC_x86_64_unknown_linux_gnu=x86_64-linux-gnu-gcc
 ENV CXX_x86_64_unknown_linux_gnu=x86_64-linux-gnu-g++
+ENV CARGO_TARGET_DIR=/usr/src/rustfs/target/docker-build
 
 WORKDIR /usr/src/rustfs
 
@@ -112,27 +123,36 @@ ENV CARGO_NET_GIT_FETCH_WITH_CLI=true \
 # Generate protobuf/flatbuffers code (uses protoc/flatc from distro)
 RUN --mount=type=cache,target=/usr/local/cargo/registry \
     --mount=type=cache,target=/usr/local/cargo/git \
-    --mount=type=cache,target=/usr/src/rustfs/target \
+    --mount=type=cache,target=/usr/src/rustfs/target/docker-build \
     cargo run --bin gproto
 
 # Build RustFS (target depends on TARGETPLATFORM)
 RUN --mount=type=cache,target=/usr/local/cargo/registry \
     --mount=type=cache,target=/usr/local/cargo/git \
-    --mount=type=cache,target=/usr/src/rustfs/target \
+    --mount=type=cache,target=/usr/src/rustfs/target/docker-build \
     set -eux; \
-    case "${TARGETPLATFORM:-linux/amd64}" in \
+    rustup target add x86_64-unknown-linux-gnu aarch64-unknown-linux-gnu; \
+    target_platform="${TARGETPLATFORM:-}"; \
+    if [ -z "${target_platform}" ]; then \
+    case "${TARGETARCH:-$(uname -m)}" in \
+    amd64|x86_64) target_platform="linux/amd64" ;; \
+    arm64|aarch64) target_platform="linux/arm64" ;; \
+    *) echo "Unsupported target architecture: ${TARGETARCH:-$(uname -m)}" >&2; exit 1 ;; \
+    esac; \
+    fi; \
+    case "${target_platform}" in \
     linux/amd64) \
     echo "Building for x86_64-unknown-linux-gnu"; \
     cargo build --release --locked --target x86_64-unknown-linux-gnu --bin rustfs -j "$(nproc)"; \
-    install -m 0755 target/x86_64-unknown-linux-gnu/release/rustfs /usr/local/bin/rustfs \
+    install -m 0755 "${CARGO_TARGET_DIR}/x86_64-unknown-linux-gnu/release/rustfs" /usr/local/bin/rustfs \
     ;; \
     linux/arm64) \
     echo "Building for aarch64-unknown-linux-gnu"; \
     cargo build --release --locked --target aarch64-unknown-linux-gnu --bin rustfs -j "$(nproc)"; \
-    install -m 0755 target/aarch64-unknown-linux-gnu/release/rustfs /usr/local/bin/rustfs \
+    install -m 0755 "${CARGO_TARGET_DIR}/aarch64-unknown-linux-gnu/release/rustfs" /usr/local/bin/rustfs \
     ;; \
     *) \
-    echo "Unsupported TARGETPLATFORM=${TARGETPLATFORM}" >&2; exit 1 \
+    echo "Unsupported target platform=${target_platform}" >&2; exit 1 \
     ;; \
     esac
 
@@ -178,7 +198,7 @@ CMD ["cargo", "run", "--bin", "rustfs", "--"]
 # -----------------------------
 # Runtime stage (Ubuntu minimal)
 # -----------------------------
-FROM ubuntu:22.04
+FROM ubuntu:24.04
 
 ARG BUILD_DATE
 ARG VCS_REF
@@ -195,6 +215,7 @@ RUN set -eux; \
     apt-get update; \
     apt-get install -y --no-install-recommends \
     ca-certificates \
+    curl \
     tzdata \
     coreutils; \
     rm -rf /var/lib/apt/lists/*

README.md

@@ -115,7 +115,7 @@ chown -R 10001:10001 data logs
 docker run -d -p 9000:9000 -p 9001:9001 -v $(pwd)/data:/data -v $(pwd)/logs:/logs rustfs/rustfs:latest
 
 # Using specific version
-docker run -d -p 9000:9000 -p 9001:9001 -v $(pwd)/data:/data -v $(pwd)/logs:/logs rustfs/rustfs:1.0.0-alpha.76
+docker run -d -p 9000:9000 -p 9001:9001 -v $(pwd)/data:/data -v $(pwd)/logs:/logs rustfs/rustfs:1.0.0-beta.4
 ```
 
 If you use [podman](https://github.com/containers/podman) instead of docker, you can install the RustFS with the below command
@@ -136,6 +136,23 @@ Similarly, you can run the command with podman
 podman compose --profile observability up -d
 ```
 
+Webhook notification quick start (Docker):
+
+```bash
+docker run -d --name rustfs -p 9000:9000 \
+  -e RUSTFS_NOTIFY_ENABLE=true \
+  -e RUSTFS_NOTIFY_WEBHOOK_ENABLE_PRIMARY=on \
+  -e RUSTFS_NOTIFY_WEBHOOK_ENDPOINT_PRIMARY=http://<host-ip>:3020/webhook \
+  -e RUSTFS_NOTIFY_WEBHOOK_QUEUE_DIR_PRIMARY=/tmp/rustfs-events \
+  rustfs/rustfs:latest
+```
+
+Notes:
+- `RUSTFS_NOTIFY_ENABLE=true` enables the global notify module switch.
+- For ARN `arn:rustfs:sqs::primary:webhook`, use instance-scoped env vars with `_PRIMARY`.
+- If queue dir is omitted, default is `/opt/rustfs/events`; ensure it is writable by the container runtime user.
+- `RUSTFS_NOTIFY_WEBHOOK_SKIP_TLS_VERIFY_PRIMARY` defaults to `false`; enabling it skips webhook TLS certificate verification, allows MITM attacks, and emits a startup warning. Prefer `RUSTFS_NOTIFY_WEBHOOK_CLIENT_CA_PRIMARY` for private CAs.
+
 **NOTE**: We recommend reviewing the `docker-compose.yml` file before running. It defines several services including Grafana, Prometheus, and Jaeger, which are helpful for RustFS observability. If you wish to start Redis or Nginx containers, you can specify the corresponding profiles.
 
 ### 3\. Build from Source (Option 3) - Advanced Users

README_ZH.md

@@ -1,4 +1,4 @@
-[![RustFS](https://github.com/user-attachments/assets/1b5afcd6-a2c3-47ff-8bc3-ce882b0ddca7)](https://rustfs.com.cn)
+[![RustFS](https://repository-images.githubusercontent.com/722597620/0fa936a2-8164-4f53-867f-def4beb64b21)](https://rustfs.com.cn)
 
 <p align="center">RustFS 是一个基于 Rust 构建的高性能分布式对象存储系统。</p>
 
@@ -113,7 +113,7 @@ RustFS 容器以非 root 用户 `rustfs` (UID `10001`) 运行。如果您使用
  docker run -d -p 9000:9000 -p 9001:9001 -v $(pwd)/data:/data -v $(pwd)/logs:/logs rustfs/rustfs:latest
 
  # 使用指定版本运行
- docker run -d -p 9000:9000 -p 9001:9001 -v $(pwd)/data:/data -v $(pwd)/logs:/logs rustfs/rustfs:1.0.0.alpha.68
+ docker run -d -p 9000:9000 -p 9001:9001 -v $(pwd)/data:/data -v $(pwd)/logs:/logs rustfs/rustfs:1.0.0-beta.4
 ```
 
 您也可以使用 Docker Compose。使用根目录下的 `docker-compose.yml` 文件：

build-rustfs.sh

@@ -109,6 +109,7 @@ FORCE_CONSOLE_UPDATE=false
 CONSOLE_VERSION="latest"
 SKIP_VERIFICATION=false
 CUSTOM_PLATFORM=""
+FEATURES=""
 
 # Print usage
 usage() {
@@ -141,6 +142,7 @@ usage() {
     echo "  --no-console               Skip console static assets"
     echo "  --force-console-update     Force update console assets even if they exist"
     echo "  --console-version VERSION  Console version to download (default: latest)"
+    echo "  -f, --features FEATURES    Cargo features to enable (e.g. 'webdav', 'full')"
     echo "  --skip-verification        Skip binary verification after build"
     echo "  -h, --help                 Show this help message"
     echo ""
@@ -148,6 +150,8 @@ usage() {
     echo "  $0                         # Build for current platform (includes console assets)"
     echo "  $0 --dev                   # Development build"
     echo "  $0 --sign                  # Build and sign binary (release CI)"
+    echo "  $0 --features webdav       # Build with WebDAV support"
+    echo "  $0 --features full         # Build with all protocol features"
     echo "  $0 --no-console            # Build without console static assets"
     echo "  $0 --force-console-update  # Force update console assets"
     echo "  $0 --platform x86_64-unknown-linux-musl  # Build for specific platform"
@@ -213,7 +217,7 @@ setup_rust_environment() {
     # Set up environment variables for musl targets
     if [[ "$PLATFORM" == *"musl"* ]]; then
         print_message $YELLOW "Setting up environment for musl target..."
-        export RUSTFLAGS="'--cfg tokio_unstable -C target-feature=-crt-static'"
+        export RUSTFLAGS="--cfg tokio_unstable -C target-feature=-crt-static"
 
         # For cargo-zigbuild, set up additional environment variables
         if command -v cargo-zigbuild &> /dev/null; then
@@ -440,6 +444,12 @@ build_binary() {
     build_cmd+=" --target $PLATFORM"
     build_cmd+=" -p rustfs --bins"
 
+    if [ -n "$FEATURES" ]; then
+        local quoted_features
+        printf -v quoted_features '%q' "$FEATURES"
+        build_cmd+=" --features $quoted_features"
+    fi
+
     print_message $BLUE "📦 Executing: $build_cmd"
 
     # Execute build (this matches exactly what the working version does)
@@ -497,6 +507,9 @@ build_rustfs() {
         print_message $YELLOW "   Force Console Update: $FORCE_CONSOLE_UPDATE"
     fi
     print_message $YELLOW "   Skip Verification: $SKIP_VERIFICATION"
+    if [ -n "$FEATURES" ]; then
+        print_message $YELLOW "   Features: $FEATURES"
+    fi
     echo ""
 
     # Setup environment
@@ -565,6 +578,15 @@ while [[ $# -gt 0 ]]; do
             SKIP_VERIFICATION=true
             shift
             ;;
+        -f|--features)
+            if [ -z "${2:-}" ]; then
+                print_message $RED "❌ Missing value for $1"
+                usage
+                exit 1
+            fi
+            FEATURES="$2"
+            shift 2
+            ;;
         -h|--help)
             usage
             exit 0

crates/appauth/README.md

@@ -1,37 +0,0 @@
-[![RustFS](https://rustfs.com/images/rustfs-github.png)](https://rustfs.com)
-
-# RustFS AppAuth - Application Authentication
-
-<p align="center">
-  <strong>Application-level authentication and authorization module for RustFS distributed object storage</strong>
-</p>
-
-<p align="center">
-  <a href="https://github.com/rustfs/rustfs/actions/workflows/ci.yml"><img alt="CI" src="https://github.com/rustfs/rustfs/actions/workflows/ci.yml/badge.svg" /></a>
-  <a href="https://docs.rustfs.com/">📖 Documentation</a>
-  · <a href="https://github.com/rustfs/rustfs/issues">🐛 Bug Reports</a>
-  · <a href="https://github.com/rustfs/rustfs/discussions">💬 Discussions</a>
-</p>
-
----
-
-## 📖 Overview
-
-**RustFS AppAuth** provides application-level authentication and authorization capabilities for the [RustFS](https://rustfs.com) distributed object storage system. For the complete RustFS experience, please visit the [main RustFS repository](https://github.com/rustfs/rustfs).
-
-## ✨ Features
-
-- JWT-based authentication with secure token management
-- RBAC (Role-Based Access Control) for fine-grained permissions
-- Multi-tenant application isolation and management
-- OAuth 2.0 and OpenID Connect integration
-- API key management and rotation
-- Session management with configurable expiration
-
-## 📚 Documentation
-
-For comprehensive documentation, examples, and usage guides, please visit the main [RustFS repository](https://github.com/rustfs/rustfs).
-
-## 📄 License
-
-This project is licensed under the Apache License 2.0 - see the [LICENSE](../../LICENSE) file for details.

crates/appauth/src/token.rs

@@ -1,128 +0,0 @@
-// Copyright 2024 RustFS Team
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use rsa::{
-    Pkcs1v15Encrypt, RsaPrivateKey, RsaPublicKey,
-    pkcs8::{DecodePrivateKey, DecodePublicKey},
-};
-use serde::{Deserialize, Serialize};
-use std::io::{Error, Result};
-
-#[derive(Serialize, Deserialize, Debug, Default, Clone)]
-pub struct Token {
-    pub name: String, // Application ID
-    pub expired: u64, // Expiry time (UNIX timestamp)
-}
-
-/// Public key generation Token
-/// [token] Token object
-/// [key] Public key string
-/// Returns the encrypted string processed by base64
-pub fn gencode(token: &Token, key: &str) -> Result<String> {
-    let data = serde_json::to_vec(token)?;
-    let mut rng = rand::rng();
-    let public_key = RsaPublicKey::from_public_key_pem(key).map_err(Error::other)?;
-    let encrypted_data = public_key.encrypt(&mut rng, Pkcs1v15Encrypt, &data).map_err(Error::other)?;
-    Ok(base64_simd::URL_SAFE_NO_PAD.encode_to_string(&encrypted_data))
-}
-
-/// Private key resolution Token
-/// [token] Encrypted string processed by base64
-/// [key] Private key string
-/// Return to the Token object
-pub fn parse(token: &str, key: &str) -> Result<Token> {
-    let encrypted_data = base64_simd::URL_SAFE_NO_PAD
-        .decode_to_vec(token.as_bytes())
-        .map_err(Error::other)?;
-    let private_key = RsaPrivateKey::from_pkcs8_pem(key).map_err(Error::other)?;
-    let decrypted_data = private_key.decrypt(Pkcs1v15Encrypt, &encrypted_data).map_err(Error::other)?;
-    let res: Token = serde_json::from_slice(&decrypted_data)?;
-    Ok(res)
-}
-
-pub fn parse_license(license: &str) -> Result<Token> {
-    parse(license, TEST_PRIVATE_KEY)
-    // match parse(license, TEST_PRIVATE_KEY) {
-    //     Ok(token) => {
-    //         if token.expired > SystemTime::now().duration_since(UNIX_EPOCH)?.as_secs() {
-    //             Ok(token)
-    //         } else {
-    //             Err("Token expired".into())
-    //         }
-    //     }
-    //     Err(e) => Err(e),
-    // }
-}
-
-static TEST_PRIVATE_KEY: &str = "-----BEGIN PRIVATE KEY-----\nMIIEvAIBADANBgkqhkiG9w0BAQEFAASCBKYwggSiAgEAAoIBAQCj86SrJIuxSxR6\nBJ/dlJEUIj6NeBRnhLQlCDdovuz61+7kJXVcxaR66w4m8W7SLEUP+IlPtnn6vmiG\n7XMhGNHIr7r1JsEVVLhZmL3tKI66DEZl786ZhG81BWqUlmcooIPS8UEPZNqJXLuz\nVGhxNyVGbj/tV7QC2pSISnKaixc+nrhxvo7w56p5qrm9tik0PjTgfZsUePkoBsSN\npoRkAauS14MAzK6HGB75CzG3dZqXUNWSWVocoWtQbZUwFGXyzU01ammsHQDvc2xu\nK1RQpd1qYH5bOWZ0N0aPFwT0r59HztFXg9sbjsnuhO1A7OiUOkc6iGVuJ0wm/9nA\nwZIBqzgjAgMBAAECggEAPMpeSEbotPhNw2BrllE76ec4omPfzPJbiU+em+wPGoNu\nRJHPDnMKJbl6Kd5jZPKdOOrCnxfd6qcnQsBQa/kz7+GYxMV12l7ra+1Cnujm4v0i\nLTHZvPpp8ZLsjeOmpF3AAzsJEJgon74OqtOlVjVIUPEYKvzV9ijt4gsYq0zfdYv0\nhrTMzyrGM4/UvKLsFIBROAfCeWfA7sXLGH8JhrRAyDrtCPzGtyyAmzoHKHtHafcB\nuyPFw/IP8otAgpDk5iiQPNkH0WwzAQIm12oHuNUa66NwUK4WEjXTnDg8KeWLHHNv\nIfN8vdbZchMUpMIvvkr7is315d8f2cHCB5gEO+GWAQKBgQDR/0xNll+FYaiUKCPZ\nvkOCAd3l5mRhsqnjPQ/6Ul1lAyYWpoJSFMrGGn/WKTa/FVFJRTGbBjwP+Mx10bfb\ngUg2GILDTISUh54fp4zngvTi9w4MWGKXrb7I1jPkM3vbJfC/v2fraQ/r7qHPpO2L\nf6ZbGxasIlSvr37KeGoelwcAQQKBgQDH3hmOTS2Hl6D4EXdq5meHKrfeoicGN7m8\noQK7u8iwn1R9zK5nh6IXxBhKYNXNwdCQtBZVRvFjjZ56SZJb7lKqa1BcTsgJfZCy\nnI3Uu4UykrECAH8AVCVqBXUDJmeA2yE+gDAtYEjvhSDHpUfWxoGHr0B/Oqk2Lxc/\npRy1qV5fYwKBgBWSL/hYVf+RhIuTg/s9/BlCr9SJ0g3nGGRrRVTlWQqjRCpXeFOO\nJzYqSq9pFGKUggEQxoOyJEFPwVDo9gXqRcyov+Xn2kaXl7qQr3yoixc1YZALFDWY\nd1ySBEqQr0xXnV9U/gvEgwotPRnjSzNlLWV2ZuHPtPtG/7M0o1H5GZMBAoGAKr3N\nW0gX53o+my4pCnxRQW+aOIsWq1a5aqRIEFudFGBOUkS2Oz+fI1P1GdrRfhnnfzpz\n2DK+plp/vIkFOpGhrf4bBlJ2psjqa7fdANRFLMaAAfyXLDvScHTQTCcnVUAHQPVq\n2BlSH56pnugyj7SNuLV6pnql+wdhAmRN2m9o1h8CgYAbX2juSr4ioXwnYjOUdrIY\n4+ERvHcXdjoJmmPcAm4y5NbSqLXyU0FQmplNMt2A5LlniWVJ9KNdjAQUt60FZw/+\nr76LdxXaHNZghyx0BOs7mtq5unSQXamZ8KixasfhE9uz3ij1jXjG6hafWkS8/68I\nuWbaZqgvy7a9oPHYlKH7Jg==\n-----END PRIVATE KEY-----\n";
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-    use rsa::{
-        RsaPrivateKey,
-        pkcs8::{EncodePrivateKey, EncodePublicKey, LineEnding},
-    };
-    use std::time::{SystemTime, UNIX_EPOCH};
-
-    #[test]
-    fn test_gencode_and_parse() {
-        let mut rng = rand::rng();
-        let bits = 2048;
-        let private_key = RsaPrivateKey::new(&mut rng, bits).expect("Failed to generate private key");
-        let public_key = RsaPublicKey::from(&private_key);
-
-        let private_key_pem = private_key.to_pkcs8_pem(LineEnding::LF).unwrap();
-        let public_key_pem = public_key.to_public_key_pem(LineEnding::LF).unwrap();
-
-        let token = Token {
-            name: "test_app".to_string(),
-            expired: SystemTime::now().duration_since(UNIX_EPOCH).unwrap().as_secs() + 3600, // 1 hour from now
-        };
-
-        let encoded = gencode(&token, &public_key_pem).expect("Failed to encode token");
-
-        let decoded = parse(&encoded, &private_key_pem).expect("Failed to decode token");
-
-        assert_eq!(token.name, decoded.name);
-        assert_eq!(token.expired, decoded.expired);
-    }
-
-    #[test]
-    fn test_parse_invalid_token() {
-        let mut rng = rand::rng();
-        let private_key_pem = RsaPrivateKey::new(&mut rng, 2048)
-            .expect("Failed to generate private key")
-            .to_pkcs8_pem(LineEnding::LF)
-            .unwrap();
-
-        let invalid_token = "invalid_base64_token";
-        let result = parse(invalid_token, &private_key_pem);
-
-        assert!(result.is_err());
-    }
-
-    #[test]
-    fn test_gencode_with_invalid_key() {
-        let token = Token {
-            name: "test_app".to_string(),
-            expired: SystemTime::now().duration_since(UNIX_EPOCH).unwrap().as_secs() + 3600, // 1 hour from now
-        };
-
-        let invalid_key = "invalid_public_key";
-        let result = gencode(&token, invalid_key);
-
-        assert!(result.is_err());
-    }
-}

crates/audit/AGENTS.md

@@ -0,0 +1,53 @@
+# Audit Crate Instructions
+
+Applies to `crates/audit/`.
+
+`rustfs-audit` is the domain layer for audit event fan-out and observability.
+It composes shared plugin/runtime abstractions from `rustfs-targets` and keeps
+audit-specific dispatch semantics, state transitions, and metrics in this
+crate.
+
+## Domain Boundaries
+
+- Keep audit-specific behavior here:
+  - audit event shaping and fan-out pipeline
+  - audit system lifecycle/state transitions
+  - audit metrics and reporting
+- Keep shared plugin/runtime mechanics in `rustfs-targets`:
+  - no duplicated replay worker orchestration
+  - no duplicated runtime manager primitives
+  - no plugin install/control-plane modeling in this crate
+
+## Runtime Layering Rules
+
+- `pipeline.rs` hosts:
+  - `AuditPipeline` (dispatch and snapshot access)
+  - `AuditRuntimeFacade` (runtime mutation path)
+  - `AuditRuntimeView` (runtime read path)
+- `registry.rs` should remain the single owner of runtime target container and
+  plugin registry composition for audit.
+- `system.rs` should coordinate lifecycle by calling facade/view/registry
+  boundaries rather than embedding low-level runtime logic.
+
+## Change Style
+
+- Preserve audit delivery semantics and error handling behavior unless the task
+  explicitly changes them.
+- Prefer extending shared abstractions in `rustfs-targets` over patching
+  one-off audit-only runtime flows.
+- Keep logging and observability machine-meaningful; avoid noisy churn in hot
+  dispatch paths.
+
+## Testing
+
+- Keep unit tests close to changed modules.
+- Keep pipeline-layer regressions in `tests/pipeline_layer_test.rs`.
+- Add regression tests for:
+  - runtime facade activation/replace/stop/shutdown behavior
+  - runtime view target/snapshot access
+  - system reload and runtime commit/clear boundaries
+- Suggested validation:
+  - `cargo test -p rustfs-audit`
+  - Focused: `cargo test -p rustfs-audit --test pipeline_layer_test`
+  - Focused: `cargo test -p rustfs-audit pipeline`
+- Full gate before commit: `make pre-commit`

crates/audit/Cargo.toml

@@ -29,8 +29,7 @@ categories = ["web-programming", "development-tools", "asynchronous", "api-bindi
 rustfs-targets = { workspace = true }
 rustfs-config = { workspace = true, features = ["audit", "constants"] }
 rustfs-ecstore = { workspace = true }
-rustfs-s3-common = { workspace = true }
-async-trait = { workspace = true }
+rustfs-s3-types = { workspace = true }
 chrono = { workspace = true }
 const-str = { workspace = true }
 futures = { workspace = true }
@@ -41,15 +40,14 @@ serde_json = { workspace = true }
 thiserror = { workspace = true }
 tokio = { workspace = true, features = ["sync", "fs", "rt-multi-thread", "rt", "time", "macros"] }
 tracing = { workspace = true, features = ["std", "attributes"] }
-url = { workspace = true }
-rumqttc = { workspace = true }
 
 [dev-dependencies]
+async-trait = { workspace = true }
 temp-env = { workspace = true }
+url = { workspace = true }
 
 [lints]
 workspace = true
 
 [lib]
-test = false
 doctest = false

crates/audit/src/entity.rs

@@ -14,7 +14,7 @@
 
 use chrono::{DateTime, Utc};
 use hashbrown::HashMap;
-use rustfs_s3_common::EventName;
+use rustfs_s3_types::EventName;
 use serde::{Deserialize, Serialize};
 use serde_json::Value;
 

crates/audit/src/factory.rs

@@ -13,109 +13,65 @@
 // limitations under the License.
 
 use crate::AuditEntry;
-use async_trait::async_trait;
-use rustfs_config::AUDIT_DEFAULT_DIR;
-use rustfs_config::audit::{AUDIT_MQTT_KEYS, AUDIT_NATS_KEYS, AUDIT_PULSAR_KEYS, AUDIT_WEBHOOK_KEYS};
-use rustfs_ecstore::config::KVS;
-use rustfs_targets::{
-    Target,
-    config::{
-        build_mqtt_args, build_nats_args, build_pulsar_args, build_webhook_args, validate_mqtt_config, validate_nats_config,
-        validate_pulsar_config, validate_webhook_config,
-    },
-    error::TargetError,
-    target::TargetType,
-};
-use std::collections::HashSet;
+use rustfs_targets::catalog::builtin::builtin_audit_target_descriptors;
+use rustfs_targets::{BuiltinTargetDescriptor, TargetPluginDescriptor};
 
-/// Trait for creating targets from configuration
-#[async_trait]
-pub trait TargetFactory: Send + Sync {
-    /// Creates a target from configuration
-    async fn create_target(&self, id: String, config: &KVS) -> Result<Box<dyn Target<AuditEntry> + Send + Sync>, TargetError>;
-
-    /// Validates target configuration
-    fn validate_config(&self, id: &str, config: &KVS) -> Result<(), TargetError>;
-
-    /// Returns a set of valid configuration field names for this target type.
-    /// This is used to filter environment variables.
-    fn get_valid_fields(&self) -> HashSet<String>;
+pub fn builtin_target_descriptors() -> Vec<BuiltinTargetDescriptor<AuditEntry>> {
+    builtin_audit_target_descriptors::<AuditEntry>()
 }
 
-/// Factory for creating Webhook targets
-pub struct WebhookTargetFactory;
-
-#[async_trait]
-impl TargetFactory for WebhookTargetFactory {
-    async fn create_target(&self, id: String, config: &KVS) -> Result<Box<dyn Target<AuditEntry> + Send + Sync>, TargetError> {
-        let args = build_webhook_args(config, AUDIT_DEFAULT_DIR, TargetType::AuditLog)?;
-        let target = rustfs_targets::target::webhook::WebhookTarget::new(id, args)?;
-        Ok(Box::new(target))
-    }
-
-    fn validate_config(&self, _id: &str, config: &KVS) -> Result<(), TargetError> {
-        validate_webhook_config(config, AUDIT_DEFAULT_DIR)
-    }
-
-    fn get_valid_fields(&self) -> HashSet<String> {
-        AUDIT_WEBHOOK_KEYS.iter().map(|s| s.to_string()).collect()
-    }
+pub fn builtin_target_plugins() -> Vec<TargetPluginDescriptor<AuditEntry>> {
+    builtin_target_descriptors()
+        .into_iter()
+        .map(|descriptor| descriptor.plugin().clone())
+        .collect()
 }
 
-/// Factory for creating MQTT targets
-pub struct MQTTTargetFactory;
+#[cfg(test)]
+mod tests {
+    use super::builtin_target_descriptors;
+    use rustfs_config::audit::AUDIT_AMQP_KEYS;
+    use rustfs_config::{AMQP_EXCHANGE, AMQP_QUEUE_DIR, AMQP_ROUTING_KEY, AMQP_URL};
+    use rustfs_ecstore::config::KVS;
+    use rustfs_targets::target::ChannelTargetType;
 
-#[async_trait]
-impl TargetFactory for MQTTTargetFactory {
-    async fn create_target(&self, id: String, config: &KVS) -> Result<Box<dyn Target<AuditEntry> + Send + Sync>, TargetError> {
-        let args = build_mqtt_args(config, AUDIT_DEFAULT_DIR, TargetType::AuditLog)?;
-        let target = rustfs_targets::target::mqtt::MQTTTarget::new(id, args)?;
-        Ok(Box::new(target))
+    fn amqp_base_config() -> KVS {
+        let mut config = KVS::new();
+        config.insert(AMQP_URL.to_string(), "amqp://127.0.0.1:5672/%2f".to_string());
+        config.insert(AMQP_EXCHANGE.to_string(), "rustfs.audit".to_string());
+        config.insert(AMQP_ROUTING_KEY.to_string(), "audit".to_string());
+        config.insert(AMQP_QUEUE_DIR.to_string(), String::new());
+        config
     }
 
-    fn validate_config(&self, _id: &str, config: &KVS) -> Result<(), TargetError> {
-        validate_mqtt_config(config)
+    #[test]
+    fn builtin_plugins_include_amqp_descriptor() {
+        let plugin = builtin_target_descriptors()
+            .into_iter()
+            .find(|plugin| plugin.plugin().target_type() == ChannelTargetType::Amqp.as_str())
+            .expect("amqp plugin should exist");
+
+        assert!(plugin.plugin().valid_fields().contains(&AMQP_URL));
+        assert!(plugin.plugin().valid_fields().contains(&AMQP_EXCHANGE));
+        assert!(plugin.plugin().valid_fields().contains(&AMQP_ROUTING_KEY));
+        assert_eq!(plugin.plugin().valid_fields().len(), AUDIT_AMQP_KEYS.len());
     }
 
-    fn get_valid_fields(&self) -> HashSet<String> {
-        AUDIT_MQTT_KEYS.iter().map(|s| s.to_string()).collect()
-    }
-}
-
-pub struct NATSTargetFactory;
-
-#[async_trait]
-impl TargetFactory for NATSTargetFactory {
-    async fn create_target(&self, id: String, config: &KVS) -> Result<Box<dyn Target<AuditEntry> + Send + Sync>, TargetError> {
-        let args = build_nats_args(config, AUDIT_DEFAULT_DIR, TargetType::AuditLog)?;
-        let target = rustfs_targets::target::nats::NATSTarget::new(id, args)?;
-        Ok(Box::new(target))
-    }
-
-    fn validate_config(&self, _id: &str, config: &KVS) -> Result<(), TargetError> {
-        validate_nats_config(config, AUDIT_DEFAULT_DIR)
-    }
-
-    fn get_valid_fields(&self) -> HashSet<String> {
-        AUDIT_NATS_KEYS.iter().map(|s| s.to_string()).collect()
-    }
-}
-
-pub struct PulsarTargetFactory;
-
-#[async_trait]
-impl TargetFactory for PulsarTargetFactory {
-    async fn create_target(&self, id: String, config: &KVS) -> Result<Box<dyn Target<AuditEntry> + Send + Sync>, TargetError> {
-        let args = build_pulsar_args(config, AUDIT_DEFAULT_DIR, TargetType::AuditLog)?;
-        let target = rustfs_targets::target::pulsar::PulsarTarget::new(id, args)?;
-        Ok(Box::new(target))
-    }
-
-    fn validate_config(&self, _id: &str, config: &KVS) -> Result<(), TargetError> {
-        validate_pulsar_config(config, AUDIT_DEFAULT_DIR)
-    }
-
-    fn get_valid_fields(&self) -> HashSet<String> {
-        AUDIT_PULSAR_KEYS.iter().map(|s| s.to_string()).collect()
+    #[test]
+    fn builtin_plugins_create_audit_amqp_target() {
+        let plugin = builtin_target_descriptors()
+            .into_iter()
+            .find(|plugin| plugin.plugin().target_type() == ChannelTargetType::Amqp.as_str())
+            .expect("amqp plugin should exist");
+
+        let target = plugin
+            .plugin()
+            .create_target("primary".to_string(), &amqp_base_config())
+            .expect("AMQP audit target should be created");
+
+        let target_id = target.id();
+        assert_eq!(target_id.id, "primary");
+        assert_eq!(target_id.name, "amqp");
+        assert!(target.store().is_none());
     }
 }

crates/audit/src/lib.rs

@@ -23,6 +23,7 @@ pub mod error;
 pub mod factory;
 pub mod global;
 pub mod observability;
+pub mod pipeline;
 pub mod registry;
 pub mod system;
 
@@ -30,5 +31,6 @@ pub use entity::{ApiDetails, AuditEntry, ObjectVersion};
 pub use error::{AuditError, AuditResult};
 pub use global::*;
 pub use observability::{AuditMetrics, AuditMetricsReport, PerformanceValidation};
+pub use pipeline::{AuditPipeline, AuditRuntimeFacade, AuditRuntimeView};
 pub use registry::AuditRegistry;
 pub use system::{AuditSystem, AuditTargetMetricSnapshot};

crates/audit/src/pipeline.rs

@@ -0,0 +1,355 @@
+//  Copyright 2024 RustFS Team
+//
+//  Licensed under the Apache License, Version 2.0 (the "License");
+//  you may not use this file except in compliance with the License.
+//  You may obtain a copy of the License at
+//
+//      http://www.apache.org/licenses/LICENSE-2.0
+//
+//  Unless required by applicable law or agreed to in writing, software
+//  distributed under the License is distributed on an "AS IS" BASIS,
+//  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+//  See the License for the specific language governing permissions and
+//  limitations under the License.
+
+use crate::{AuditEntry, AuditResult, observability, system::AuditTargetMetricSnapshot};
+use rustfs_targets::{
+    BuiltinPluginRuntimeAdapter, PluginRuntimeAdapter, ReplayEvent, ReplayWorkerManager, RuntimeActivation, SharedTarget, Target,
+    target::EntityTarget,
+};
+use std::sync::Arc;
+use std::time::Duration;
+use tokio::sync::{Mutex, RwLock};
+use tracing::{error, info, warn};
+
+#[derive(Clone)]
+pub struct AuditPipeline {
+    registry: Arc<Mutex<crate::AuditRegistry>>,
+}
+
+impl AuditPipeline {
+    pub fn new(registry: Arc<Mutex<crate::AuditRegistry>>) -> Self {
+        Self { registry }
+    }
+
+    pub async fn dispatch(&self, entry: Arc<AuditEntry>) -> AuditResult<()> {
+        let start_time = std::time::Instant::now();
+
+        let targets: Vec<SharedTarget<AuditEntry>> = {
+            let registry = self.registry.lock().await;
+            let targets = registry.list_target_values();
+
+            if targets.is_empty() {
+                warn!("No audit targets configured for dispatch");
+                return Ok(());
+            }
+
+            targets
+        };
+
+        let mut tasks = Vec::new();
+
+        for target in targets {
+            let entity_target = EntityTarget {
+                object_name: entry.api.name.clone().unwrap_or_default(),
+                bucket_name: entry.api.bucket.clone().unwrap_or_default(),
+                event_name: entry.event,
+                data: (*entry).clone(),
+            };
+
+            let task = async move {
+                let result = target.save(Arc::new(entity_target)).await;
+                (target.id().to_string(), result)
+            };
+
+            tasks.push(task);
+        }
+
+        let results = futures::future::join_all(tasks).await;
+
+        let mut errors = Vec::new();
+        let mut success_count = 0;
+
+        for (target_key, result) in results {
+            match result {
+                Ok(_) => {
+                    success_count += 1;
+                    observability::record_target_success();
+                }
+                Err(e) => {
+                    error!(target_id = %target_key, error = %e, "Failed to dispatch audit log to target");
+                    errors.push(e);
+                    observability::record_target_failure();
+                }
+            }
+        }
+
+        let dispatch_time = start_time.elapsed();
+
+        if errors.is_empty() {
+            observability::record_audit_success(dispatch_time);
+        } else {
+            observability::record_audit_failure(dispatch_time);
+            warn!(
+                error_count = errors.len(),
+                success_count = success_count,
+                "Some audit targets failed to receive log entry"
+            );
+        }
+
+        Ok(())
+    }
+
+    pub async fn dispatch_batch(&self, entries: Vec<Arc<AuditEntry>>) -> AuditResult<()> {
+        let start_time = std::time::Instant::now();
+
+        let targets: Vec<SharedTarget<AuditEntry>> = {
+            let registry = self.registry.lock().await;
+            let targets = registry.list_target_values();
+
+            if targets.is_empty() {
+                warn!("No audit targets configured for batch dispatch");
+                return Ok(());
+            }
+
+            targets
+        };
+
+        let mut tasks = Vec::new();
+        for target in targets {
+            let entries_clone: Vec<_> = entries.iter().map(Arc::clone).collect();
+
+            let task = async move {
+                let mut success_count = 0;
+                let mut errors = Vec::new();
+                for entry in entries_clone {
+                    let entity_target = EntityTarget {
+                        object_name: entry.api.name.clone().unwrap_or_default(),
+                        bucket_name: entry.api.bucket.clone().unwrap_or_default(),
+                        event_name: entry.event,
+                        data: (*entry).clone(),
+                    };
+                    match target.save(Arc::new(entity_target)).await {
+                        Ok(_) => success_count += 1,
+                        Err(e) => errors.push(e),
+                    }
+                }
+                (target.id().to_string(), success_count, errors)
+            };
+            tasks.push(task);
+        }
+
+        let results = futures::future::join_all(tasks).await;
+        let mut total_success = 0;
+        let mut total_errors = 0;
+        for (_target_id, success_count, errors) in results {
+            total_success += success_count;
+            total_errors += errors.len();
+            for e in errors {
+                error!("Batch dispatch error: {:?}", e);
+            }
+        }
+
+        let dispatch_time = start_time.elapsed();
+        info!(
+            "Batch dispatched {} entries, success: {}, errors: {}, time: {:?}",
+            entries.len(),
+            total_success,
+            total_errors,
+            dispatch_time
+        );
+
+        Ok(())
+    }
+
+    pub async fn snapshot_target_metrics(&self) -> Vec<AuditTargetMetricSnapshot> {
+        let registry = self.registry.lock().await;
+        registry
+            .list_target_values()
+            .into_iter()
+            .map(|target| {
+                let delivery = target.delivery_snapshot();
+                AuditTargetMetricSnapshot {
+                    failed_messages: delivery.failed_messages,
+                    queue_length: delivery.queue_length,
+                    target_id: target.id().to_string(),
+                    total_messages: delivery.total_messages,
+                }
+            })
+            .collect()
+    }
+
+    pub async fn snapshot_target_health(&self) -> Vec<rustfs_targets::RuntimeTargetHealthSnapshot> {
+        let registry = self.registry.lock().await;
+        registry.runtime_manager().health_snapshots().await
+    }
+}
+
+#[derive(Clone)]
+pub struct AuditRuntimeView {
+    registry: Arc<Mutex<crate::AuditRegistry>>,
+}
+
+impl AuditRuntimeView {
+    pub fn new(registry: Arc<Mutex<crate::AuditRegistry>>) -> Self {
+        Self { registry }
+    }
+
+    pub async fn list_targets(&self) -> Vec<String> {
+        let registry = self.registry.lock().await;
+        registry.list_targets()
+    }
+
+    pub async fn get_target_values(&self) -> Vec<rustfs_targets::SharedTarget<AuditEntry>> {
+        let registry = self.registry.lock().await;
+        registry.list_target_values()
+    }
+
+    pub async fn get_target(&self, target_id: &str) -> Option<String> {
+        let registry = self.registry.lock().await;
+        registry.get_target(target_id).map(|target| target.id().to_string())
+    }
+
+    pub async fn enable_target(&self, target_id: &str) -> AuditResult<()> {
+        let registry = self.registry.lock().await;
+        if registry.get_target(target_id).is_some() {
+            info!(target_id = %target_id, "Target enabled");
+            Ok(())
+        } else {
+            Err(crate::AuditError::Configuration(format!("Target not found: {target_id}"), None))
+        }
+    }
+
+    pub async fn disable_target(&self, target_id: &str) -> AuditResult<()> {
+        let registry = self.registry.lock().await;
+        if registry.get_target(target_id).is_some() {
+            info!(target_id = %target_id, "Target disabled");
+            Ok(())
+        } else {
+            Err(crate::AuditError::Configuration(format!("Target not found: {target_id}"), None))
+        }
+    }
+
+    pub async fn remove_target(&self, target_id: &str) -> AuditResult<()> {
+        let mut registry = self.registry.lock().await;
+        if registry.remove_target(target_id).await.is_some() {
+            info!(target_id = %target_id, "Target removed");
+            Ok(())
+        } else {
+            Err(crate::AuditError::Configuration(format!("Target not found: {target_id}"), None))
+        }
+    }
+
+    pub async fn upsert_target(&self, target_id: String, target: Box<dyn Target<AuditEntry> + Send + Sync>) -> AuditResult<()> {
+        if let Err(err) = target.init().await {
+            return Err(crate::AuditError::Target(err));
+        }
+
+        let shared_target: SharedTarget<AuditEntry> = Arc::from(target);
+        let mut registry = self.registry.lock().await;
+        let _ = registry.remove_target(&target_id).await;
+        registry.add_shared_target(target_id.clone(), shared_target);
+        info!(target_id = %target_id, "Target upserted");
+        Ok(())
+    }
+}
+
+#[derive(Clone)]
+pub struct AuditRuntimeFacade {
+    registry: Arc<Mutex<crate::AuditRegistry>>,
+    replay_workers: Arc<RwLock<ReplayWorkerManager>>,
+    runtime_adapter: Arc<dyn PluginRuntimeAdapter<AuditEntry>>,
+}
+
+impl AuditRuntimeFacade {
+    pub fn new(registry: Arc<Mutex<crate::AuditRegistry>>, replay_workers: Arc<RwLock<ReplayWorkerManager>>) -> Self {
+        let runtime_adapter = BuiltinPluginRuntimeAdapter::new(
+            Arc::new(move |event: ReplayEvent<AuditEntry>| {
+                Box::pin(async move {
+                    match event {
+                        ReplayEvent::Delivered { key, target } => {
+                            info!("Successfully sent audit entry, target: {}, key: {}", target.id(), key.to_string());
+                            observability::record_target_success();
+                        }
+                        ReplayEvent::RetryableError { error, target, .. } => match error {
+                            rustfs_targets::TargetError::NotConnected => {
+                                warn!("Target {} not connected, retrying...", target.id());
+                            }
+                            rustfs_targets::TargetError::Timeout(_) => {
+                                warn!("Timeout sending to target {}, retrying...", target.id());
+                            }
+                            _ => {}
+                        },
+                        ReplayEvent::Dropped { reason, target, .. } => {
+                            warn!("Dropped queued payload for target {}: {}", target.id(), reason);
+                            observability::record_target_failure();
+                        }
+                        ReplayEvent::PermanentFailure { error, target, .. } => {
+                            error!("Permanent error for target {}: {}", target.id(), error);
+                            target.record_final_failure();
+                            observability::record_target_failure();
+                        }
+                        ReplayEvent::RetryExhausted { key, target } => {
+                            warn!("Max retries exceeded for key {}, target: {}, skipping", key.to_string(), target.id());
+                            target.record_final_failure();
+                            observability::record_target_failure();
+                        }
+                        ReplayEvent::UnreadableEntry { key, error, target } => {
+                            warn!("Skipping unreadable audit store entry {} for target {}: {}", key, target.id(), error);
+                        }
+                    }
+                })
+            }),
+            Arc::new(|target_id, has_replay| {
+                if has_replay {
+                    info!(target_id = %target_id, "Audit stream processing started");
+                } else {
+                    info!(target_id = %target_id, "No store configured, skip audit stream processing");
+                }
+            }),
+            None,
+            Duration::from_millis(500),
+            Duration::from_millis(500),
+            "Stopping audit stream",
+        );
+
+        Self {
+            registry,
+            replay_workers,
+            runtime_adapter: Arc::new(runtime_adapter),
+        }
+    }
+
+    pub async fn replace_targets(&self, activation: RuntimeActivation<AuditEntry>) -> AuditResult<()> {
+        let mut registry = self.registry.lock().await;
+        let mut replay_workers = self.replay_workers.write().await;
+        self.runtime_adapter
+            .replace_runtime_targets(registry.runtime_manager_mut(), &mut replay_workers, activation)
+            .await
+            .map_err(crate::AuditError::Target)?;
+        Ok(())
+    }
+
+    pub async fn shutdown_runtime(
+        &self,
+        registry: &mut crate::AuditRegistry,
+        replay_workers: &mut ReplayWorkerManager,
+    ) -> AuditResult<()> {
+        self.runtime_adapter
+            .shutdown(registry.runtime_manager_mut(), replay_workers)
+            .await
+            .map_err(crate::AuditError::Target)
+    }
+
+    pub async fn activate_targets_with_replay(
+        &self,
+        targets: Vec<Box<dyn Target<AuditEntry> + Send + Sync>>,
+    ) -> RuntimeActivation<AuditEntry> {
+        self.runtime_adapter.activate_with_replay(targets).await
+    }
+
+    pub async fn stop_replay_workers(&self) {
+        let mut replay_workers = self.replay_workers.write().await;
+        self.runtime_adapter.stop_replay_workers(&mut replay_workers).await;
+    }
+}

crates/audit/src/registry.rs

@@ -12,26 +12,19 @@
 //  See the License for the specific language governing permissions and
 //  limitations under the License.
 
-use crate::{
-    AuditEntry, AuditError, AuditResult,
-    factory::{MQTTTargetFactory, NATSTargetFactory, PulsarTargetFactory, TargetFactory, WebhookTargetFactory},
-};
-use futures::StreamExt;
-use futures::stream::FuturesUnordered;
-use hashbrown::HashMap;
+use crate::{AuditEntry, AuditError, AuditResult, factory::builtin_target_plugins};
 use rustfs_config::audit::AUDIT_ROUTE_PREFIX;
 use rustfs_ecstore::config::{Config, KVS};
 use rustfs_targets::arn::TargetID;
-use rustfs_targets::{Target, TargetError, config::collect_target_configs, target::ChannelTargetType};
-use std::sync::Arc;
-use tracing::{error, info};
+use rustfs_targets::{SharedTarget, Target, TargetError, TargetPluginRegistry, TargetRuntimeManager};
+use tracing::info;
 
 /// Registry for managing audit targets
 pub struct AuditRegistry {
     /// Storage for created targets
-    targets: HashMap<String, Box<dyn Target<AuditEntry> + Send + Sync>>,
-    /// Factories for creating targets
-    factories: HashMap<String, Box<dyn TargetFactory>>,
+    targets: TargetRuntimeManager<AuditEntry>,
+    /// Registered plugins for creating targets
+    plugins: TargetPluginRegistry<AuditEntry>,
 }
 
 impl Default for AuditRegistry {
@@ -43,27 +36,17 @@ impl Default for AuditRegistry {
 impl AuditRegistry {
     /// Creates a new AuditRegistry
     pub fn new() -> Self {
-        let mut registry = AuditRegistry {
-            factories: HashMap::new(),
-            targets: HashMap::new(),
-        };
+        let mut plugins = TargetPluginRegistry::new();
+        plugins.register_all(builtin_target_plugins());
 
-        // Register built-in factories
-        registry.register(ChannelTargetType::Webhook.as_str(), Box::new(WebhookTargetFactory));
-        registry.register(ChannelTargetType::Mqtt.as_str(), Box::new(MQTTTargetFactory));
-        registry.register(ChannelTargetType::Nats.as_str(), Box::new(NATSTargetFactory));
-        registry.register(ChannelTargetType::Pulsar.as_str(), Box::new(PulsarTargetFactory));
-
-        registry
+        AuditRegistry {
+            targets: TargetRuntimeManager::new(),
+            plugins,
+        }
     }
 
-    /// Registers a new factory for a target type
-    ///
-    /// # Arguments
-    /// * `target_type` - The type of the target (e.g., "webhook", "mqtt").
-    /// * `factory` - The factory instance to create targets of this type.
-    pub fn register(&mut self, target_type: &str, factory: Box<dyn TargetFactory>) {
-        self.factories.insert(target_type.to_string(), factory);
+    pub fn supports_target_type(&self, target_type: &str) -> bool {
+        self.plugins.supports_target_type(target_type)
     }
 
     /// Creates a target of the specified type with the given ID and configuration
@@ -81,16 +64,7 @@ impl AuditRegistry {
         id: String,
         config: &KVS,
     ) -> Result<Box<dyn Target<AuditEntry> + Send + Sync>, TargetError> {
-        let factory = self
-            .factories
-            .get(target_type)
-            .ok_or_else(|| TargetError::Configuration(format!("Unknown target type: {target_type}")))?;
-
-        // Validate configuration before creating target
-        factory.validate_config(&id, config)?;
-
-        // Create target
-        factory.create_target(id, config).await
+        self.plugins.create_target(target_type, id, config)
     }
 
     /// Creates all targets from a configuration
@@ -106,37 +80,10 @@ impl AuditRegistry {
         &self,
         config: &Config,
     ) -> AuditResult<Vec<Box<dyn Target<AuditEntry> + Send + Sync>>> {
-        let mut tasks = FuturesUnordered::new();
-        for (target_type, factory) in &self.factories {
-            tracing::Span::current().record("target_type", target_type.as_str());
-            info!("Start working on target types...");
-            let valid_fields = factory.get_valid_fields();
-            for (id, merged_config) in collect_target_configs(config, AUDIT_ROUTE_PREFIX, target_type, &valid_fields) {
-                info!(instance_id = %id, "Target is enabled, ready to create a task");
-                let tid = id.clone();
-                let merged_config_arc = Arc::new(merged_config);
-                tasks.push(async move {
-                    let result = factory.create_target(tid.clone(), &merged_config_arc).await;
-                    (tid, result)
-                });
-            }
-        }
-
-        let mut successful_targets = Vec::new();
-        while let Some((id, result)) = tasks.next().await {
-            match result {
-                Ok(target) => {
-                    info!(target_type = %target.id().name, instance_id = %id, "Create a target successfully");
-                    successful_targets.push(target);
-                }
-                Err(e) => {
-                    error!(instance_id = %id, error = %e, "Failed to create a target");
-                }
-            }
-        }
-
-        info!(count = successful_targets.len(), "All target processing completed");
-        Ok(successful_targets)
+        self.plugins
+            .create_targets_from_config(config, AUDIT_ROUTE_PREFIX)
+            .await
+            .map_err(AuditError::from)
     }
 
     /// Adds a target to the registry
@@ -144,8 +91,14 @@ impl AuditRegistry {
     /// # Arguments
     /// * `id` - The identifier for the target.
     /// * `target` - The target instance to be added.
-    pub fn add_target(&mut self, id: String, target: Box<dyn Target<AuditEntry> + Send + Sync>) {
-        self.targets.insert(id, target);
+    pub fn add_target(&mut self, _id: String, target: Box<dyn Target<AuditEntry> + Send + Sync>) {
+        debug_assert_eq!(_id, target.id().to_string());
+        self.targets.add_boxed(target);
+    }
+
+    pub fn add_shared_target(&mut self, _id: String, target: SharedTarget<AuditEntry>) {
+        debug_assert_eq!(_id, target.id().to_string());
+        self.targets.add_arc(target);
     }
 
     /// Removes a target from the registry
@@ -155,8 +108,8 @@ impl AuditRegistry {
     ///
     /// # Returns
     /// * `Option<Box<dyn Target<AuditEntry> + Send + Sync>>` - The removed target if it existed.
-    pub fn remove_target(&mut self, id: &str) -> Option<Box<dyn Target<AuditEntry> + Send + Sync>> {
-        self.targets.remove(id)
+    pub async fn remove_target(&mut self, id: &str) -> Option<rustfs_targets::SharedTarget<AuditEntry>> {
+        self.targets.remove_and_close(id).await
     }
 
     /// Gets a target from the registry
@@ -166,13 +119,21 @@ impl AuditRegistry {
     ///
     /// # Returns
     /// * `Option<&(dyn Target<AuditEntry> + Send + Sync)>` - The target if it exists.
-    pub fn get_target(&self, id: &str) -> Option<&(dyn Target<AuditEntry> + Send + Sync)> {
-        self.targets.get(id).map(|t| t.as_ref())
+    pub fn get_target(&self, id: &str) -> Option<rustfs_targets::SharedTarget<AuditEntry>> {
+        self.targets.get(id)
     }
 
     /// Lists cloned target values for runtime inspection without exposing mutable registry access.
-    pub fn list_target_values(&self) -> Vec<Box<dyn Target<AuditEntry> + Send + Sync>> {
-        self.targets.values().map(|target| target.clone_dyn()).collect()
+    pub fn list_target_values(&self) -> Vec<rustfs_targets::SharedTarget<AuditEntry>> {
+        self.targets.values()
+    }
+
+    pub fn runtime_manager(&self) -> &TargetRuntimeManager<AuditEntry> {
+        &self.targets
+    }
+
+    pub fn runtime_manager_mut(&mut self) -> &mut TargetRuntimeManager<AuditEntry> {
+        &mut self.targets
     }
 
     /// Lists all target IDs
@@ -180,7 +141,7 @@ impl AuditRegistry {
     /// # Returns
     /// * `Vec<String>` - A vector of all target IDs in the registry.
     pub fn list_targets(&self) -> Vec<String> {
-        self.targets.keys().cloned().collect()
+        self.targets.keys()
     }
 
     /// Closes all targets and clears the registry
@@ -188,20 +149,23 @@ impl AuditRegistry {
     /// # Returns
     /// * `AuditResult<()>` - Result indicating success or failure.
     pub async fn close_all(&mut self) -> AuditResult<()> {
-        let mut errors = Vec::new();
+        let mut first_error = None;
 
-        for (id, target) in self.targets.drain() {
-            if let Err(e) = target.close().await {
-                error!(target_id = %id, error = %e, "Failed to close audit target");
-                errors.push(e);
+        for target_id in self.targets.keys() {
+            if let Some(target) = self.targets.remove(&target_id)
+                && let Err(err) = target.close().await
+            {
+                tracing::error!(target_id = %target_id, error = %err, "Failed to close target during shutdown");
+                if first_error.is_none() {
+                    first_error = Some(err);
+                }
             }
         }
 
-        if !errors.is_empty() {
-            return Err(AuditError::Target(errors.into_iter().next().unwrap()));
+        match first_error {
+            Some(err) => Err(AuditError::Target(err)),
+            None => Ok(()),
         }
-
-        Ok(())
     }
 
     /// Creates a unique key for a target based on its type and ID
@@ -276,7 +240,107 @@ impl AuditRegistry {
         target: Box<dyn Target<AuditEntry> + Send + Sync>,
     ) -> AuditResult<()> {
         let key = self.create_key(target_type, target_id);
-        self.targets.insert(key, target);
+        debug_assert_eq!(key, target.id().to_string());
+        self.targets.add_boxed(target);
         Ok(())
     }
 }
+
+#[cfg(test)]
+mod tests {
+    use super::AuditRegistry;
+    use crate::{AuditEntry, AuditError};
+    use rustfs_targets::arn::TargetID;
+    use rustfs_targets::store::{Key, Store};
+    use rustfs_targets::target::{ChannelTargetType, EntityTarget, QueuedPayload, QueuedPayloadMeta};
+    use rustfs_targets::{StoreError, Target, TargetError};
+    use std::sync::Arc;
+    use std::sync::atomic::{AtomicUsize, Ordering};
+
+    #[derive(Clone)]
+    struct CloseTestTarget {
+        id: TargetID,
+        close_calls: Arc<AtomicUsize>,
+        fail_on_close: bool,
+    }
+
+    impl CloseTestTarget {
+        fn new(id: TargetID, close_calls: Arc<AtomicUsize>, fail_on_close: bool) -> Self {
+            Self {
+                id,
+                close_calls,
+                fail_on_close,
+            }
+        }
+    }
+
+    #[async_trait::async_trait]
+    impl Target<AuditEntry> for CloseTestTarget {
+        fn id(&self) -> TargetID {
+            self.id.clone()
+        }
+
+        async fn is_active(&self) -> Result<bool, TargetError> {
+            Ok(true)
+        }
+
+        async fn save(&self, _event: Arc<EntityTarget<AuditEntry>>) -> Result<(), TargetError> {
+            Ok(())
+        }
+
+        async fn send_raw_from_store(&self, _key: Key, _body: Vec<u8>, _meta: QueuedPayloadMeta) -> Result<(), TargetError> {
+            Ok(())
+        }
+
+        async fn close(&self) -> Result<(), TargetError> {
+            self.close_calls.fetch_add(1, Ordering::SeqCst);
+            if self.fail_on_close {
+                Err(TargetError::Unknown("close failed".to_string()))
+            } else {
+                Ok(())
+            }
+        }
+
+        fn store(&self) -> Option<&(dyn Store<QueuedPayload, Error = StoreError, Key = Key> + Send + Sync)> {
+            None
+        }
+
+        fn clone_dyn(&self) -> Box<dyn Target<AuditEntry> + Send + Sync> {
+            Box::new(self.clone())
+        }
+
+        fn is_enabled(&self) -> bool {
+            true
+        }
+    }
+
+    #[test]
+    fn registry_registers_amqp_factory() {
+        let registry = AuditRegistry::new();
+
+        assert!(registry.supports_target_type(ChannelTargetType::Amqp.as_str()));
+    }
+
+    #[tokio::test]
+    async fn close_all_returns_first_error_and_clears_targets() {
+        let mut registry = AuditRegistry::new();
+        let ok_calls = Arc::new(AtomicUsize::new(0));
+        let fail_calls = Arc::new(AtomicUsize::new(0));
+
+        let ok_id = TargetID::new("ok".to_string(), "webhook".to_string());
+        let fail_id = TargetID::new("fail".to_string(), "webhook".to_string());
+
+        registry.add_target(ok_id.to_string(), Box::new(CloseTestTarget::new(ok_id, Arc::clone(&ok_calls), false)));
+        registry.add_target(
+            fail_id.to_string(),
+            Box::new(CloseTestTarget::new(fail_id, Arc::clone(&fail_calls), true)),
+        );
+
+        let result = registry.close_all().await;
+
+        assert!(matches!(result, Err(AuditError::Target(TargetError::Unknown(_)))));
+        assert_eq!(ok_calls.load(Ordering::SeqCst), 1);
+        assert_eq!(fail_calls.load(Ordering::SeqCst), 1);
+        assert!(registry.list_targets().is_empty());
+    }
+}

crates/audit/src/system.rs

@@ -12,16 +12,14 @@
 //  See the License for the specific language governing permissions and
 //  limitations under the License.
 
-use crate::{AuditEntry, AuditError, AuditRegistry, AuditResult, observability};
-use hashbrown::HashMap;
-use rustfs_ecstore::config::Config;
-use rustfs_targets::{
-    StoreError, Target, TargetError,
-    store::{Key, Store},
-    target::{EntityTarget, QueuedPayload},
+use crate::{
+    AuditEntry, AuditError, AuditRegistry, AuditResult, observability,
+    pipeline::{AuditPipeline, AuditRuntimeFacade, AuditRuntimeView},
 };
+use rustfs_ecstore::config::Config;
+use rustfs_targets::{ReplayWorkerManager, Target};
 use std::sync::Arc;
-use tokio::sync::{Mutex, RwLock, mpsc};
+use tokio::sync::{Mutex, RwLock};
 use tracing::{error, info, warn};
 
 #[derive(Debug, Clone, Default, PartialEq, Eq)]
@@ -49,7 +47,7 @@ pub struct AuditSystem {
     state: Arc<RwLock<AuditSystemState>>,
     config: Arc<RwLock<Option<Config>>>,
     /// Cancellation senders for active audit stream tasks (target_id -> cancel tx)
-    stream_cancellers: Arc<RwLock<HashMap<String, mpsc::Sender<()>>>>,
+    stream_cancellers: Arc<RwLock<ReplayWorkerManager>>,
 }
 
 impl Default for AuditSystem {
@@ -59,16 +57,68 @@ impl Default for AuditSystem {
 }
 
 impl AuditSystem {
+    fn pipeline(&self) -> AuditPipeline {
+        AuditPipeline::new(self.registry.clone())
+    }
+
+    fn runtime_view(&self) -> AuditRuntimeView {
+        AuditRuntimeView::new(self.registry.clone())
+    }
+
+    fn runtime_facade(&self) -> AuditRuntimeFacade {
+        AuditRuntimeFacade::new(self.registry.clone(), self.stream_cancellers.clone())
+    }
+
     /// Creates a new audit system
     pub fn new() -> Self {
         Self {
             registry: Arc::new(Mutex::new(AuditRegistry::new())),
             state: Arc::new(RwLock::new(AuditSystemState::Stopped)),
             config: Arc::new(RwLock::new(None)),
-            stream_cancellers: Arc::new(RwLock::new(HashMap::new())),
+            stream_cancellers: Arc::new(RwLock::new(ReplayWorkerManager::new())),
         }
     }
 
+    async fn create_targets_from_config(&self, config: &Config) -> AuditResult<Vec<Box<dyn Target<AuditEntry> + Send + Sync>>> {
+        let registry = self.registry.lock().await;
+        registry.create_audit_targets_from_config(config).await
+    }
+
+    async fn clear_runtime_targets(&self) -> AuditResult<()> {
+        {
+            let mut registry = self.registry.lock().await;
+            let mut replay_workers = self.stream_cancellers.write().await;
+            self.runtime_facade()
+                .shutdown_runtime(&mut registry, &mut replay_workers)
+                .await?;
+        }
+
+        let mut state = self.state.write().await;
+        *state = AuditSystemState::Stopped;
+        Ok(())
+    }
+
+    async fn commit_runtime_targets(
+        &self,
+        targets: Vec<Box<dyn Target<AuditEntry> + Send + Sync>>,
+        final_state: AuditSystemState,
+    ) -> AuditResult<()> {
+        if targets.is_empty() {
+            info!("No enabled audit targets found, keeping audit system stopped");
+            self.clear_runtime_targets().await?;
+            return Ok(());
+        }
+
+        info!(target_count = targets.len(), "Created audit targets successfully");
+
+        let activation = self.runtime_facade().activate_targets_with_replay(targets).await;
+        self.runtime_facade().replace_targets(activation).await?;
+
+        let mut state = self.state.write().await;
+        *state = final_state;
+        Ok(())
+    }
+
     /// Starts the audit system with the given configuration
     ///
     /// # Arguments
@@ -103,31 +153,14 @@ impl AuditSystem {
             *config_guard = Some(config.clone());
         }
 
-        // Create targets from configuration
-        let mut registry = self.registry.lock().await;
-        match registry.create_audit_targets_from_config(&config).await {
+        match self.create_targets_from_config(&config).await {
             Ok(targets) => {
-                if targets.is_empty() {
-                    info!("No enabled audit targets found, keeping audit system stopped");
-                    drop(registry);
-                    return Ok(());
-                }
-
                 {
                     let mut state = self.state.write().await;
                     *state = AuditSystemState::Starting;
                 }
 
-                info!(target_count = targets.len(), "Created audit targets successfully");
-
-                // Initialize all targets
-                for target in targets {
-                    self.init_and_register_target(target, &mut registry).await;
-                }
-
-                // Update state to running
-                let mut state = self.state.write().await;
-                *state = AuditSystemState::Running;
+                self.commit_runtime_targets(targets, AuditSystemState::Running).await?;
                 info!("Audit system started successfully");
                 Ok(())
             }
@@ -207,18 +240,10 @@ impl AuditSystem {
         info!("Stopping audit system");
 
         // Stop all stream tasks first
-        self.stop_all_streams().await;
-
-        // Close all targets
-        let mut registry = self.registry.lock().await;
-        if let Err(e) = registry.close_all().await {
+        if let Err(e) = self.clear_runtime_targets().await {
             error!(error = %e, "Failed to close some audit targets");
         }
 
-        // Update state to stopped
-        let mut state = self.state.write().await;
-        *state = AuditSystemState::Stopped;
-
         // Clear configuration
         let mut config_guard = self.config.write().await;
         *config_guard = None;
@@ -248,8 +273,6 @@ impl AuditSystem {
     /// # Returns
     /// * `AuditResult<()>` - Result indicating success or failure
     pub async fn dispatch(&self, entry: Arc<AuditEntry>) -> AuditResult<()> {
-        let start_time = std::time::Instant::now();
-
         let state = self.state.read().await;
 
         match *state {
@@ -262,77 +285,7 @@ impl AuditSystem {
             }
         }
         drop(state);
-
-        // Collect cloned targets under lock, then dispatch without holding it
-        let targets: Vec<(String, Box<dyn Target<AuditEntry> + Send + Sync>)> = {
-            let registry = self.registry.lock().await;
-            let target_keys = registry.list_targets();
-
-            if target_keys.is_empty() {
-                warn!("No audit targets configured for dispatch");
-                return Ok(());
-            }
-
-            target_keys
-                .into_iter()
-                .filter_map(|key| registry.get_target(&key).map(|t| (key, t.clone_dyn())))
-                .collect()
-        };
-
-        // Dispatch to all targets concurrently (no lock held)
-        let mut tasks = Vec::new();
-
-        for (target_key, target) in targets {
-            let entity_target = EntityTarget {
-                object_name: entry.api.name.clone().unwrap_or_default(),
-                bucket_name: entry.api.bucket.clone().unwrap_or_default(),
-                event_name: entry.event,
-                data: (*entry).clone(),
-            };
-
-            let task = async move {
-                let result = target.save(Arc::new(entity_target)).await;
-                (target_key, result)
-            };
-
-            tasks.push(task);
-        }
-
-        // Execute all dispatch tasks
-        let results = futures::future::join_all(tasks).await;
-
-        let mut errors = Vec::new();
-        let mut success_count = 0;
-
-        for (target_key, result) in results {
-            match result {
-                Ok(_) => {
-                    success_count += 1;
-                    observability::record_target_success();
-                }
-                Err(e) => {
-                    error!(target_id = %target_key, error = %e, "Failed to dispatch audit log to target");
-                    errors.push(e);
-                    observability::record_target_failure();
-                }
-            }
-        }
-
-        let dispatch_time = start_time.elapsed();
-
-        if errors.is_empty() {
-            observability::record_audit_success(dispatch_time);
-        } else {
-            observability::record_audit_failure(dispatch_time);
-            // Log errors but don't fail the entire dispatch
-            warn!(
-                error_count = errors.len(),
-                success_count = success_count,
-                "Some audit targets failed to receive log entry"
-            );
-        }
-
-        Ok(())
+        self.pipeline().dispatch(entry).await
     }
 
     /// Dispatches a batch of audit log entries to all active targets
@@ -343,239 +296,12 @@ impl AuditSystem {
     /// # Returns
     /// * `AuditResult<()>` - Result indicating success or failure
     pub async fn dispatch_batch(&self, entries: Vec<Arc<AuditEntry>>) -> AuditResult<()> {
-        let start_time = std::time::Instant::now();
-
         let state = self.state.read().await;
         if *state != AuditSystemState::Running {
             return Err(AuditError::NotInitialized("Audit system is not running".to_string()));
         }
         drop(state);
-
-        // Collect targets under lock, then dispatch without holding it
-        let targets: Vec<(String, Box<dyn Target<AuditEntry> + Send + Sync>)> = {
-            let registry = self.registry.lock().await;
-            let target_keys = registry.list_targets();
-
-            if target_keys.is_empty() {
-                warn!("No audit targets configured for batch dispatch");
-                return Ok(());
-            }
-
-            target_keys
-                .into_iter()
-                .filter_map(|key| registry.get_target(&key).map(|t| (key, t.clone_dyn())))
-                .collect()
-        };
-
-        let mut tasks = Vec::new();
-        for (target_key, target) in targets {
-            let entries_clone: Vec<_> = entries.iter().map(Arc::clone).collect();
-            let target_key_clone = target_key.clone();
-
-            let task = async move {
-                let mut success_count = 0;
-                let mut errors = Vec::new();
-                for entry in entries_clone {
-                    let entity_target = EntityTarget {
-                        object_name: entry.api.name.clone().unwrap_or_default(),
-                        bucket_name: entry.api.bucket.clone().unwrap_or_default(),
-                        event_name: entry.event,
-                        data: (*entry).clone(),
-                    };
-                    match target.save(Arc::new(entity_target)).await {
-                        Ok(_) => success_count += 1,
-                        Err(e) => errors.push(e),
-                    }
-                }
-                (target_key_clone, success_count, errors)
-            };
-            tasks.push(task);
-        }
-
-        let results = futures::future::join_all(tasks).await;
-        let mut total_success = 0;
-        let mut total_errors = 0;
-        for (_target_id, success_count, errors) in results {
-            total_success += success_count;
-            total_errors += errors.len();
-            for e in errors {
-                error!("Batch dispatch error: {:?}", e);
-            }
-        }
-
-        let dispatch_time = start_time.elapsed();
-        info!(
-            "Batch dispatched {} entries, success: {}, errors: {}, time: {:?}",
-            entries.len(),
-            total_success,
-            total_errors,
-            dispatch_time
-        );
-
-        Ok(())
-    }
-
-    /// Stops all active audit stream tasks by sending cancellation signals.
-    async fn stop_all_streams(&self) {
-        let mut cancellers = self.stream_cancellers.write().await;
-        for (target_id, cancel_tx) in cancellers.drain() {
-            info!(target_id = %target_id, "Stopping audit stream");
-            let _ = cancel_tx.send(()).await;
-        }
-    }
-
-    /// Initializes a single target: runs init(), starts stream if store is present,
-    /// and adds it to the registry. For store-backed targets, registration and stream
-    /// startup proceed even if init() fails so queued entries can be drained later.
-    async fn init_and_register_target(
-        &self,
-        target: Box<dyn Target<AuditEntry> + Send + Sync>,
-        registry: &mut AuditRegistry,
-    ) -> Option<String> {
-        let target_id = target.id().to_string();
-        let has_store = target.store().is_some();
-
-        if let Err(e) = target.init().await {
-            error!(target_id = %target_id, error = %e, "Failed to initialize audit target");
-            // Non-store targets: init failure is fatal.
-            if !has_store {
-                return None;
-            }
-            // Store-backed targets: still register and start the stream so queued
-            // entries can be drained when connectivity recovers.
-            warn!(
-                target_id = %target_id,
-                "Proceeding with store-backed audit target despite init failure"
-            );
-        }
-
-        if target.is_enabled() {
-            if let Some(store) = target.store() {
-                info!(target_id = %target_id, "Start audit stream processing for target");
-                let store_clone: Box<dyn Store<QueuedPayload, Error = StoreError, Key = Key> + Send> = store.boxed_clone();
-                let target_arc: Arc<dyn Target<AuditEntry> + Send + Sync> = Arc::from(target.clone_dyn());
-                let cancel_tx = self.start_audit_stream_with_batching(store_clone, target_arc);
-
-                self.stream_cancellers.write().await.insert(target_id.clone(), cancel_tx);
-                info!(target_id = %target_id, "Audit stream processing started");
-            } else {
-                info!(target_id = %target_id, "No store configured, skip audit stream processing");
-            }
-        } else {
-            info!(target_id = %target_id, "Target disabled, skip audit stream processing");
-        }
-
-        registry.add_target(target_id.clone(), target);
-        Some(target_id)
-    }
-
-    /// Starts the audit stream processing for a target with batching and retry logic
-    ///
-    /// # Arguments
-    /// * `store` - The store from which to read audit entries
-    /// * `target` - The target to which audit entries will be sent
-    ///
-    /// This function spawns a background task that continuously reads audit entries from the provided store
-    /// and attempts to send them to the specified target. It implements retry logic with exponential backoff
-    fn start_audit_stream_with_batching(
-        &self,
-        store: Box<dyn Store<QueuedPayload, Error = StoreError, Key = Key> + Send>,
-        target: Arc<dyn Target<AuditEntry> + Send + Sync>,
-    ) -> mpsc::Sender<()> {
-        let (cancel_tx, mut cancel_rx) = mpsc::channel(1);
-        let state = self.state.clone();
-
-        tokio::spawn(async move {
-            use std::time::Duration;
-            use tokio::time::sleep;
-
-            info!("Starting audit stream for target: {}", target.id());
-
-            const MAX_RETRIES: usize = 5;
-            const BASE_RETRY_DELAY: Duration = Duration::from_secs(2);
-
-            loop {
-                // Check for cancellation signal
-                if cancel_rx.try_recv().is_ok() {
-                    info!("Audit stream cancelled for target: {}", target.id());
-                    break;
-                }
-
-                match *state.read().await {
-                    AuditSystemState::Running | AuditSystemState::Paused | AuditSystemState::Starting => {}
-                    _ => {
-                        info!("Audit stream stopped for target: {}", target.id());
-                        break;
-                    }
-                }
-
-                let keys: Vec<Key> = store.list();
-                if keys.is_empty() {
-                    tokio::select! {
-                        _ = sleep(Duration::from_millis(500)) => {},
-                        _ = cancel_rx.recv() => {
-                            info!("Audit stream cancelled during idle for target: {}", target.id());
-                            return;
-                        }
-                    }
-                    continue;
-                }
-
-                for key in keys {
-                    if cancel_rx.try_recv().is_ok() {
-                        info!("Audit stream cancelled during processing for target: {}", target.id());
-                        return;
-                    }
-
-                    let mut retries = 0usize;
-                    let mut success = false;
-
-                    while retries < MAX_RETRIES && !success {
-                        match target.send_from_store(key.clone()).await {
-                            Ok(_) => {
-                                info!("Successfully sent audit entry, target: {}, key: {}", target.id(), key.to_string());
-                                observability::record_target_success();
-                                success = true;
-                            }
-                            Err(e) => {
-                                match &e {
-                                    TargetError::NotConnected => {
-                                        warn!("Target {} not connected, retrying...", target.id());
-                                    }
-                                    TargetError::Timeout(_) => {
-                                        warn!("Timeout sending to target {}, retrying...", target.id());
-                                    }
-                                    TargetError::Dropped(reason) => {
-                                        warn!("Dropped queued payload for target {}: {}", target.id(), reason);
-                                        observability::record_target_failure();
-                                        break;
-                                    }
-                                    _ => {
-                                        error!("Permanent error for target {}: {}", target.id(), e);
-                                        target.record_final_failure();
-                                        observability::record_target_failure();
-                                        break;
-                                    }
-                                }
-                                retries += 1;
-                                let backoff = BASE_RETRY_DELAY * (1 << retries);
-                                sleep(backoff).await;
-                            }
-                        }
-                    }
-
-                    if retries >= MAX_RETRIES && !success {
-                        warn!("Max retries exceeded for key {}, target: {}, skipping", key.to_string(), target.id());
-                        target.record_final_failure();
-                        observability::record_target_failure();
-                    }
-                }
-
-                sleep(Duration::from_millis(100)).await;
-            }
-        });
-
-        cancel_tx
+        self.pipeline().dispatch_batch(entries).await
     }
 
     /// Enables a specific target
@@ -586,15 +312,7 @@ impl AuditSystem {
     /// # Returns
     /// * `AuditResult<()>` - Result indicating success or failure
     pub async fn enable_target(&self, target_id: &str) -> AuditResult<()> {
-        // This would require storing enabled/disabled state per target
-        // For now, just check if target exists
-        let registry = self.registry.lock().await;
-        if registry.get_target(target_id).is_some() {
-            info!(target_id = %target_id, "Target enabled");
-            Ok(())
-        } else {
-            Err(AuditError::Configuration(format!("Target not found: {target_id}"), None))
-        }
+        self.runtime_view().enable_target(target_id).await
     }
 
     /// Disables a specific target
@@ -605,15 +323,7 @@ impl AuditSystem {
     /// # Returns
     /// * `AuditResult<()>` - Result indicating success or failure
     pub async fn disable_target(&self, target_id: &str) -> AuditResult<()> {
-        // This would require storing enabled/disabled state per target
-        // For now, just check if target exists
-        let registry = self.registry.lock().await;
-        if registry.get_target(target_id).is_some() {
-            info!(target_id = %target_id, "Target disabled");
-            Ok(())
-        } else {
-            Err(AuditError::Configuration(format!("Target not found: {target_id}"), None))
-        }
+        self.runtime_view().disable_target(target_id).await
     }
 
     /// Removes a target from the system
@@ -624,16 +334,7 @@ impl AuditSystem {
     /// # Returns
     /// * `AuditResult<()>` - Result indicating success or failure
     pub async fn remove_target(&self, target_id: &str) -> AuditResult<()> {
-        let mut registry = self.registry.lock().await;
-        if let Some(target) = registry.remove_target(target_id) {
-            if let Err(e) = target.close().await {
-                error!(target_id = %target_id, error = %e, "Failed to close removed target");
-            }
-            info!(target_id = %target_id, "Target removed");
-            Ok(())
-        } else {
-            Err(AuditError::Configuration(format!("Target not found: {target_id}"), None))
-        }
+        self.runtime_view().remove_target(target_id).await
     }
 
     /// Updates or inserts a target
@@ -645,23 +346,7 @@ impl AuditSystem {
     /// # Returns
     /// * `AuditResult<()>` - Result indicating success or failure
     pub async fn upsert_target(&self, target_id: String, target: Box<dyn Target<AuditEntry> + Send + Sync>) -> AuditResult<()> {
-        let mut registry = self.registry.lock().await;
-
-        // Initialize the target
-        if let Err(e) = target.init().await {
-            return Err(AuditError::Target(e));
-        }
-
-        // Remove existing target if present
-        if let Some(old_target) = registry.remove_target(&target_id)
-            && let Err(e) = old_target.close().await
-        {
-            error!(target_id = %target_id, error = %e, "Failed to close old target during upsert");
-        }
-
-        registry.add_target(target_id.clone(), target);
-        info!(target_id = %target_id, "Target upserted");
-        Ok(())
+        self.runtime_view().upsert_target(target_id, target).await
     }
 
     /// Lists all targets
@@ -669,33 +354,27 @@ impl AuditSystem {
     /// # Returns
     /// * `Vec<String>` - List of target IDs
     pub async fn list_targets(&self) -> Vec<String> {
-        let registry = self.registry.lock().await;
-        registry.list_targets()
+        self.runtime_view().list_targets().await
     }
 
     /// Returns cloned target values for read-only runtime inspection.
-    pub async fn get_target_values(&self) -> Vec<Box<dyn Target<AuditEntry> + Send + Sync>> {
-        let registry = self.registry.lock().await;
-        registry.list_target_values()
+    pub async fn get_target_values(&self) -> Vec<rustfs_targets::SharedTarget<AuditEntry>> {
+        self.runtime_view().get_target_values().await
     }
 
     /// Returns per-target delivery metrics for Prometheus collection.
     pub async fn snapshot_target_metrics(&self) -> Vec<AuditTargetMetricSnapshot> {
-        let targets = self.get_target_values().await;
-        let mut snapshots = Vec::with_capacity(targets.len());
+        self.pipeline().snapshot_target_metrics().await
+    }
 
-        for target in targets {
-            let delivery = target.delivery_snapshot();
-            snapshots.push(AuditTargetMetricSnapshot {
-                failed_messages: delivery.failed_messages,
-                queue_length: delivery.queue_length,
-                target_id: target.id().to_string(),
-                total_messages: delivery.total_messages,
-            });
-        }
+    pub async fn snapshot_target_health(&self) -> Vec<rustfs_targets::RuntimeTargetHealthSnapshot> {
+        self.pipeline().snapshot_target_health().await
+    }
 
-        snapshots.sort_by(|a, b| a.target_id.cmp(&b.target_id));
-        snapshots
+    pub async fn runtime_status_snapshot(&self) -> rustfs_targets::RuntimeStatusSnapshot {
+        let replay_workers = self.stream_cancellers.read().await;
+        let registry = self.registry.lock().await;
+        registry.runtime_manager().status_snapshot(&replay_workers)
     }
 
     /// Gets information about a specific target
@@ -706,8 +385,7 @@ impl AuditSystem {
     /// # Returns
     /// * `Option<String>` - Target ID if found
     pub async fn get_target(&self, target_id: &str) -> Option<String> {
-        let registry = self.registry.lock().await;
-        registry.get_target(target_id).map(|target| target.id().to_string())
+        self.runtime_view().get_target(target_id).await
     }
 
     /// Reloads configuration and updates targets
@@ -722,32 +400,20 @@ impl AuditSystem {
 
         observability::record_config_reload();
 
-        // Stop all existing stream tasks first
-        self.stop_all_streams().await;
-
         // Store new configuration
         {
             let mut config_guard = self.config.write().await;
             *config_guard = Some(new_config.clone());
         }
 
-        // Close all existing targets
-        let mut registry = self.registry.lock().await;
-        if let Err(e) = registry.close_all().await {
-            error!(error = %e, "Failed to close existing targets during reload");
-        }
+        let final_state = match self.get_state().await {
+            AuditSystemState::Paused => AuditSystemState::Paused,
+            _ => AuditSystemState::Running,
+        };
 
-        // Create new targets from updated configuration
-        match registry.create_audit_targets_from_config(&new_config).await {
+        match self.create_targets_from_config(&new_config).await {
             Ok(targets) => {
-                info!(target_count = targets.len(), "Reloaded audit targets successfully");
-
-                for target in targets {
-                    if let Some(target_id) = self.init_and_register_target(target, &mut registry).await {
-                        info!(target_id = %target_id, "Target initialized (reload)");
-                    }
-                }
-
+                self.commit_runtime_targets(targets, final_state).await?;
                 info!("Audit configuration reloaded successfully");
                 Ok(())
             }
@@ -779,3 +445,106 @@ impl AuditSystem {
         observability::reset_metrics().await;
     }
 }
+
+#[cfg(test)]
+mod tests {
+    use super::{AuditSystem, AuditSystemState};
+    use async_trait::async_trait;
+    use rustfs_targets::ReplayWorkerManager;
+    use rustfs_targets::arn::TargetID;
+    use rustfs_targets::store::{Key, Store};
+    use rustfs_targets::target::{EntityTarget, QueuedPayload, QueuedPayloadMeta};
+    use rustfs_targets::{StoreError, Target, TargetError};
+    use serde::{Serialize, de::DeserializeOwned};
+    use std::collections::HashMap;
+    use std::sync::Arc;
+    use std::sync::atomic::{AtomicUsize, Ordering};
+    use tokio::sync::mpsc;
+
+    #[derive(Clone)]
+    struct TestTarget {
+        close_calls: Arc<AtomicUsize>,
+        id: TargetID,
+    }
+
+    impl TestTarget {
+        fn new(id: &str, name: &str) -> Self {
+            Self {
+                close_calls: Arc::new(AtomicUsize::new(0)),
+                id: TargetID::new(id.to_string(), name.to_string()),
+            }
+        }
+    }
+
+    #[async_trait]
+    impl<E> Target<E> for TestTarget
+    where
+        E: Send + Sync + 'static + Clone + Serialize + DeserializeOwned,
+    {
+        fn id(&self) -> TargetID {
+            self.id.clone()
+        }
+
+        async fn is_active(&self) -> Result<bool, TargetError> {
+            Ok(true)
+        }
+
+        async fn save(&self, _event: Arc<EntityTarget<E>>) -> Result<(), TargetError> {
+            Ok(())
+        }
+
+        async fn send_raw_from_store(&self, _key: Key, _body: Vec<u8>, _meta: QueuedPayloadMeta) -> Result<(), TargetError> {
+            Ok(())
+        }
+
+        async fn close(&self) -> Result<(), TargetError> {
+            self.close_calls.fetch_add(1, Ordering::SeqCst);
+            Ok(())
+        }
+
+        fn store(&self) -> Option<&(dyn Store<QueuedPayload, Error = StoreError, Key = Key> + Send + Sync)> {
+            None
+        }
+
+        fn clone_dyn(&self) -> Box<dyn Target<E> + Send + Sync> {
+            Box::new(self.clone())
+        }
+
+        fn is_enabled(&self) -> bool {
+            true
+        }
+    }
+
+    #[tokio::test]
+    async fn reload_with_empty_config_stops_existing_runtime() {
+        let system = AuditSystem::new();
+        let target = TestTarget::new("primary", "webhook");
+        let close_calls = Arc::clone(&target.close_calls);
+
+        {
+            let mut registry = system.registry.lock().await;
+            registry.add_target("primary:webhook".to_string(), Box::new(target));
+        }
+        {
+            let mut state = system.state.write().await;
+            *state = AuditSystemState::Running;
+        }
+        {
+            let mut replay_workers = system.stream_cancellers.write().await;
+            let (cancel_tx, _cancel_rx) = mpsc::channel(1);
+            replay_workers.insert("primary:webhook".to_string(), cancel_tx);
+            assert_eq!(replay_workers.len(), 1);
+        }
+
+        system
+            .reload_config(rustfs_ecstore::config::Config(HashMap::new()))
+            .await
+            .expect("reload with empty config should succeed");
+
+        assert_eq!(system.get_state().await, AuditSystemState::Stopped);
+        assert!(system.list_targets().await.is_empty());
+        assert_eq!(system.runtime_status_snapshot().await, ReplayWorkerManager::new().snapshot(0));
+        assert_eq!(close_calls.load(Ordering::SeqCst), 1);
+        assert_eq!(*system.config.read().await, Some(rustfs_ecstore::config::Config(HashMap::new())));
+    }
+}

crates/audit/tests/pipeline_layer_test.rs

@@ -0,0 +1,170 @@
+//  Copyright 2024 RustFS Team
+//
+//  Licensed under the Apache License, Version 2.0 (the "License");
+//  you may not use this file except in compliance with the License.
+//  You may obtain a copy of the License at
+//
+//      http://www.apache.org/licenses/LICENSE-2.0
+//
+//  Unless required by applicable law or agreed to in writing, software
+//  distributed under the License is distributed on an "AS IS" BASIS,
+//  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+//  See the License for the specific language governing permissions and
+//  limitations under the License.
+
+use async_trait::async_trait;
+use rustfs_audit::{AuditPipeline, AuditRegistry, AuditRuntimeFacade, AuditRuntimeView};
+use rustfs_targets::arn::TargetID;
+use rustfs_targets::store::{Key, Store};
+use rustfs_targets::target::{EntityTarget, QueuedPayload, QueuedPayloadMeta};
+use rustfs_targets::{StoreError, Target, TargetError};
+use serde::{Serialize, de::DeserializeOwned};
+use std::sync::Arc;
+use std::sync::atomic::{AtomicUsize, Ordering};
+use tokio::sync::{Mutex, RwLock};
+
+#[derive(Clone)]
+struct TestTarget {
+    close_calls: Arc<AtomicUsize>,
+    id: TargetID,
+    init_calls: Arc<AtomicUsize>,
+}
+
+impl TestTarget {
+    fn new(id: &str, name: &str) -> Self {
+        Self {
+            close_calls: Arc::new(AtomicUsize::new(0)),
+            id: TargetID::new(id.to_string(), name.to_string()),
+            init_calls: Arc::new(AtomicUsize::new(0)),
+        }
+    }
+}
+
+#[async_trait]
+impl<E> Target<E> for TestTarget
+where
+    E: Send + Sync + 'static + Clone + Serialize + DeserializeOwned,
+{
+    fn id(&self) -> TargetID {
+        self.id.clone()
+    }
+
+    async fn is_active(&self) -> Result<bool, TargetError> {
+        Ok(true)
+    }
+
+    async fn save(&self, _event: Arc<EntityTarget<E>>) -> Result<(), TargetError> {
+        Ok(())
+    }
+
+    async fn send_raw_from_store(&self, _key: Key, _body: Vec<u8>, _meta: QueuedPayloadMeta) -> Result<(), TargetError> {
+        Ok(())
+    }
+
+    async fn close(&self) -> Result<(), TargetError> {
+        self.close_calls.fetch_add(1, Ordering::SeqCst);
+        Ok(())
+    }
+
+    fn store(&self) -> Option<&(dyn Store<QueuedPayload, Error = StoreError, Key = Key> + Send + Sync)> {
+        None
+    }
+
+    fn clone_dyn(&self) -> Box<dyn Target<E> + Send + Sync> {
+        Box::new(self.clone())
+    }
+
+    async fn init(&self) -> Result<(), TargetError> {
+        self.init_calls.fetch_add(1, Ordering::SeqCst);
+        Ok(())
+    }
+
+    fn is_enabled(&self) -> bool {
+        true
+    }
+}
+
+#[tokio::test]
+async fn audit_runtime_view_lists_empty_targets() {
+    let registry = Arc::new(Mutex::new(AuditRegistry::new()));
+    let runtime_view = AuditRuntimeView::new(registry);
+
+    assert!(runtime_view.list_targets().await.is_empty());
+    assert!(runtime_view.get_target_values().await.is_empty());
+    assert!(runtime_view.get_target("missing").await.is_none());
+}
+
+#[tokio::test]
+async fn audit_pipeline_reports_empty_runtime_snapshots() {
+    let registry = Arc::new(Mutex::new(AuditRegistry::new()));
+    let pipeline = AuditPipeline::new(registry);
+
+    assert!(pipeline.snapshot_target_metrics().await.is_empty());
+    assert!(pipeline.snapshot_target_health().await.is_empty());
+}
+
+#[tokio::test]
+async fn audit_runtime_facade_stops_empty_replay_workers() {
+    let registry = Arc::new(Mutex::new(AuditRegistry::new()));
+    let replay_workers = Arc::new(RwLock::new(rustfs_targets::ReplayWorkerManager::new()));
+    let facade = AuditRuntimeFacade::new(registry, replay_workers);
+
+    facade.stop_replay_workers().await;
+}
+
+#[tokio::test]
+async fn audit_runtime_facade_activates_empty_target_list() {
+    let registry = Arc::new(Mutex::new(AuditRegistry::new()));
+    let replay_workers = Arc::new(RwLock::new(rustfs_targets::ReplayWorkerManager::new()));
+    let facade = AuditRuntimeFacade::new(registry, replay_workers);
+
+    let activation = facade.activate_targets_with_replay(Vec::new()).await;
+    assert!(activation.targets.is_empty());
+    assert_eq!(activation.replay_workers.len(), 0);
+}
+
+#[tokio::test]
+async fn audit_runtime_view_upsert_and_remove_target() {
+    let registry = Arc::new(Mutex::new(AuditRegistry::new()));
+    let runtime_view = AuditRuntimeView::new(registry.clone());
+    let target = TestTarget::new("primary", "webhook");
+    let init_calls = Arc::clone(&target.init_calls);
+    let close_calls = Arc::clone(&target.close_calls);
+
+    runtime_view
+        .upsert_target("primary:webhook".to_string(), Box::new(target))
+        .await
+        .expect("upsert should succeed");
+
+    assert_eq!(runtime_view.list_targets().await, vec!["primary:webhook".to_string()]);
+    assert_eq!(init_calls.load(Ordering::SeqCst), 1);
+
+    runtime_view
+        .remove_target("primary:webhook")
+        .await
+        .expect("remove should succeed");
+
+    assert!(runtime_view.list_targets().await.is_empty());
+    assert_eq!(close_calls.load(Ordering::SeqCst), 1);
+}
+
+#[tokio::test]
+async fn audit_runtime_facade_replace_targets_commits_runtime_state() {
+    let registry = Arc::new(Mutex::new(AuditRegistry::new()));
+    let replay_workers = Arc::new(RwLock::new(rustfs_targets::ReplayWorkerManager::new()));
+    let facade = AuditRuntimeFacade::new(registry.clone(), replay_workers.clone());
+    let target = TestTarget::new("primary", "webhook");
+    let activation = rustfs_targets::RuntimeActivation {
+        replay_workers: rustfs_targets::ReplayWorkerManager::new(),
+        targets: vec![Arc::new(target) as rustfs_targets::SharedTarget<rustfs_audit::AuditEntry>],
+    };
+
+    facade
+        .replace_targets(activation)
+        .await
+        .expect("replace_targets should succeed");
+
+    let runtime_view = AuditRuntimeView::new(registry);
+    assert_eq!(runtime_view.list_targets().await, vec!["primary:webhook".to_string()]);
+    assert_eq!(replay_workers.read().await.len(), 0);
+}

crates/checksums/Cargo.toml

@@ -23,7 +23,7 @@ homepage.workspace = true
 description = "Checksum calculation and verification callbacks for HTTP request and response bodies sent by service clients generated by RustFS, ensuring data integrity and authenticity."
 keywords = ["checksum-calculation", "verification", "integrity", "authenticity", "rustfs"]
 categories = ["web-programming", "development-tools", "network-programming"]
-documentation = "https://docs.rs/rustfs-signer/latest/rustfs_checksum/"
+documentation = "https://docs.rs/rustfs-checksums/latest/rustfs_checksum/"
 
 [dependencies]
 bytes = { workspace = true }

crates/common/Cargo.toml

@@ -33,12 +33,8 @@ tonic = { workspace = true }
 uuid = { workspace = true }
 chrono = { workspace = true }
 metrics = { workspace = true }
-rustfs-madmin = { workspace = true }
-rustfs-filemeta = { workspace = true }
 serde = { workspace = true }
-path-clean = { workspace = true }
 rmp-serde = { workspace = true }
-async-trait = { workspace = true }
 s3s = { workspace = true }
 tracing = { workspace = true }
 

crates/common/src/globals.rs

@@ -17,6 +17,7 @@
 use chrono::{DateTime, Utc};
 use std::collections::HashMap;
 use std::sync::LazyLock;
+use std::sync::atomic::{AtomicU64, Ordering};
 use tokio::sync::RwLock;
 use tonic::transport::Channel;
 
@@ -27,6 +28,7 @@ pub static GLOBAL_RUSTFS_ADDR: LazyLock<RwLock<String>> = LazyLock::new(|| RwLoc
 pub static GLOBAL_CONN_MAP: LazyLock<RwLock<HashMap<String, Channel>>> = LazyLock::new(|| RwLock::new(HashMap::new()));
 pub static GLOBAL_ROOT_CERT: LazyLock<RwLock<Option<Vec<u8>>>> = LazyLock::new(|| RwLock::new(None));
 pub static GLOBAL_MTLS_IDENTITY: LazyLock<RwLock<Option<MtlsIdentityPem>>> = LazyLock::new(|| RwLock::new(None));
+pub static GLOBAL_OUTBOUND_TLS_GENERATION: LazyLock<AtomicU64> = LazyLock::new(|| AtomicU64::new(0));
 /// Global initialization time of the RustFS node.
 pub static GLOBAL_INIT_TIME: LazyLock<RwLock<Option<DateTime<Utc>>>> = LazyLock::new(|| RwLock::new(None));
 
@@ -88,6 +90,16 @@ pub async fn set_global_mtls_identity(identity: Option<MtlsIdentityPem>) {
     *GLOBAL_MTLS_IDENTITY.write().await = identity;
 }
 
+/// Set the global outbound TLS generation.
+pub fn set_global_outbound_tls_generation(generation: u64) {
+    GLOBAL_OUTBOUND_TLS_GENERATION.store(generation, Ordering::Relaxed);
+}
+
+/// Get the global outbound TLS generation.
+pub fn get_global_outbound_tls_generation() -> u64 {
+    GLOBAL_OUTBOUND_TLS_GENERATION.load(Ordering::Relaxed)
+}
+
 /// Evict a stale/dead connection from the global connection cache.
 /// This is critical for cluster recovery when a node dies unexpectedly (e.g., power-off).
 /// By removing the cached connection, subsequent requests will establish a fresh connection.

crates/common/src/heal_channel.rs

@@ -260,9 +260,17 @@ pub enum HealChannelCommand {
         response_tx: oneshot::Sender<Result<HealAdmissionResult, String>>,
     },
     /// Query heal task status
-    Query { heal_path: String, client_token: String },
+    Query {
+        heal_path: String,
+        client_token: String,
+        response_tx: oneshot::Sender<Result<HealChannelResponse, String>>,
+    },
     /// Cancel heal task
-    Cancel { heal_path: String },
+    Cancel {
+        heal_path: String,
+        client_token: String,
+        response_tx: oneshot::Sender<Result<HealChannelResponse, String>>,
+    },
 }
 
 /// Heal request from admin to ahm
@@ -379,7 +387,9 @@ fn heal_response_sender() -> &'static HealResponseSender {
 
 /// Publish a heal response to subscribers.
 pub fn publish_heal_response(response: HealChannelResponse) -> Result<(), broadcast::error::SendError<HealChannelResponse>> {
-    heal_response_sender().send(response).map(|_| ())
+    let sender = heal_response_sender();
+    let _ = sender.send(response);
+    Ok(())
 }
 
 /// Subscribe to heal responses.
@@ -405,14 +415,36 @@ pub async fn send_heal_request(request: HealChannelRequest) -> Result<(), String
     }
 }
 
+async fn receive_heal_channel_response(
+    response_rx: oneshot::Receiver<Result<HealChannelResponse, String>>,
+) -> Result<HealChannelResponse, String> {
+    response_rx
+        .await
+        .map_err(|e| format!("Failed to receive heal channel response: {e}"))?
+}
+
 /// Send heal query request
-pub async fn query_heal_status(heal_path: String, client_token: String) -> Result<(), String> {
-    send_heal_command(HealChannelCommand::Query { heal_path, client_token }).await
+pub async fn query_heal_status(heal_path: String, client_token: String) -> Result<HealChannelResponse, String> {
+    let (response_tx, response_rx) = oneshot::channel();
+    send_heal_command(HealChannelCommand::Query {
+        heal_path,
+        client_token,
+        response_tx,
+    })
+    .await?;
+    receive_heal_channel_response(response_rx).await
 }
 
 /// Send heal cancel request
-pub async fn cancel_heal_task(heal_path: String) -> Result<(), String> {
-    send_heal_command(HealChannelCommand::Cancel { heal_path }).await
+pub async fn cancel_heal_task(heal_path: String, client_token: String) -> Result<HealChannelResponse, String> {
+    let (response_tx, response_rx) = oneshot::channel();
+    send_heal_command(HealChannelCommand::Cancel {
+        heal_path,
+        client_token,
+        response_tx,
+    })
+    .await?;
+    receive_heal_channel_response(response_rx).await
 }
 
 /// Create a new heal request
@@ -626,5 +658,10 @@ mod tests {
         let received = receiver.recv().await.expect("should receive heal response");
         assert_eq!(received.request_id, response.request_id);
         assert!(received.success);
+
+        drop(receiver);
+        let response = create_heal_response("req-no-subscriber".to_string(), true, None, None);
+
+        publish_heal_response(response).expect("publish without subscribers should be ignored");
     }
 }

crates/common/src/internode_metrics.rs

@@ -1,166 +0,0 @@
-// Copyright 2024 RustFS Team
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use metrics::{counter, gauge};
-use std::sync::{
-    Arc, LazyLock,
-    atomic::{AtomicU64, Ordering},
-};
-use std::time::{Duration, SystemTime, UNIX_EPOCH};
-
-#[derive(Debug, Clone, Copy, Default, PartialEq, Eq)]
-pub struct InternodeMetricsSnapshot {
-    pub sent_bytes_total: u64,
-    pub recv_bytes_total: u64,
-    pub outgoing_requests_total: u64,
-    pub incoming_requests_total: u64,
-    pub errors_total: u64,
-    pub dial_errors_total: u64,
-    pub dial_avg_time_nanos: u64,
-    pub last_dial_unix_millis: u64,
-}
-
-#[derive(Debug, Default)]
-pub struct InternodeMetrics {
-    sent_bytes_total: AtomicU64,
-    recv_bytes_total: AtomicU64,
-    outgoing_requests_total: AtomicU64,
-    incoming_requests_total: AtomicU64,
-    errors_total: AtomicU64,
-    dial_errors_total: AtomicU64,
-    dial_total_time_nanos: AtomicU64,
-    dial_samples_total: AtomicU64,
-    last_dial_unix_millis: AtomicU64,
-}
-
-impl InternodeMetrics {
-    pub fn record_sent_bytes(&self, bytes: usize) {
-        let bytes = bytes as u64;
-        if bytes == 0 {
-            return;
-        }
-        self.sent_bytes_total.fetch_add(bytes, Ordering::Relaxed);
-        counter!("rustfs.internode.sent.bytes.total").increment(bytes);
-    }
-
-    pub fn record_recv_bytes(&self, bytes: usize) {
-        let bytes = bytes as u64;
-        if bytes == 0 {
-            return;
-        }
-        self.recv_bytes_total.fetch_add(bytes, Ordering::Relaxed);
-        counter!("rustfs.internode.recv.bytes.total").increment(bytes);
-    }
-
-    pub fn record_outgoing_request(&self) {
-        self.outgoing_requests_total.fetch_add(1, Ordering::Relaxed);
-        counter!("rustfs.internode.requests.outgoing.total").increment(1);
-    }
-
-    pub fn record_incoming_request(&self) {
-        self.incoming_requests_total.fetch_add(1, Ordering::Relaxed);
-        counter!("rustfs.internode.requests.incoming.total").increment(1);
-    }
-
-    pub fn record_error(&self) {
-        self.errors_total.fetch_add(1, Ordering::Relaxed);
-        counter!("rustfs.internode.errors.total").increment(1);
-    }
-
-    pub fn record_dial_result(&self, duration: Duration, success: bool) {
-        let elapsed_nanos = duration.as_nanos().min(u128::from(u64::MAX)) as u64;
-        self.dial_total_time_nanos.fetch_add(elapsed_nanos, Ordering::Relaxed);
-        let samples = self.dial_samples_total.fetch_add(1, Ordering::Relaxed) + 1;
-        let total = self.dial_total_time_nanos.load(Ordering::Relaxed);
-        gauge!("rustfs.internode.dial.avg_time.nanos").set(total as f64 / samples as f64);
-
-        if !success {
-            self.dial_errors_total.fetch_add(1, Ordering::Relaxed);
-            counter!("rustfs.internode.dial.errors.total").increment(1);
-        }
-
-        let now_ms = SystemTime::now()
-            .duration_since(UNIX_EPOCH)
-            .unwrap_or_default()
-            .as_millis()
-            .min(u128::from(u64::MAX)) as u64;
-        self.last_dial_unix_millis.store(now_ms, Ordering::Relaxed);
-    }
-
-    pub fn snapshot(&self) -> InternodeMetricsSnapshot {
-        let dial_samples_total = self.dial_samples_total.load(Ordering::Relaxed);
-        let dial_total_time_nanos = self.dial_total_time_nanos.load(Ordering::Relaxed);
-        let dial_avg_time_nanos = dial_total_time_nanos.checked_div(dial_samples_total).unwrap_or(0);
-
-        InternodeMetricsSnapshot {
-            sent_bytes_total: self.sent_bytes_total.load(Ordering::Relaxed),
-            recv_bytes_total: self.recv_bytes_total.load(Ordering::Relaxed),
-            outgoing_requests_total: self.outgoing_requests_total.load(Ordering::Relaxed),
-            incoming_requests_total: self.incoming_requests_total.load(Ordering::Relaxed),
-            errors_total: self.errors_total.load(Ordering::Relaxed),
-            dial_errors_total: self.dial_errors_total.load(Ordering::Relaxed),
-            dial_avg_time_nanos,
-            last_dial_unix_millis: self.last_dial_unix_millis.load(Ordering::Relaxed),
-        }
-    }
-
-    #[doc(hidden)]
-    pub fn reset_for_test(&self) {
-        self.sent_bytes_total.store(0, Ordering::Relaxed);
-        self.recv_bytes_total.store(0, Ordering::Relaxed);
-        self.outgoing_requests_total.store(0, Ordering::Relaxed);
-        self.incoming_requests_total.store(0, Ordering::Relaxed);
-        self.errors_total.store(0, Ordering::Relaxed);
-        self.dial_errors_total.store(0, Ordering::Relaxed);
-        self.dial_total_time_nanos.store(0, Ordering::Relaxed);
-        self.dial_samples_total.store(0, Ordering::Relaxed);
-        self.last_dial_unix_millis.store(0, Ordering::Relaxed);
-    }
-}
-
-pub fn global_internode_metrics() -> &'static Arc<InternodeMetrics> {
-    static GLOBAL_INTERNODE_METRICS: LazyLock<Arc<InternodeMetrics>> = LazyLock::new(|| Arc::new(InternodeMetrics::default()));
-    &GLOBAL_INTERNODE_METRICS
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-
-    #[test]
-    fn snapshot_reports_recorded_values() {
-        let metrics = global_internode_metrics();
-        metrics.reset_for_test();
-
-        metrics.record_sent_bytes(64);
-        metrics.record_recv_bytes(32);
-        metrics.record_outgoing_request();
-        metrics.record_incoming_request();
-        metrics.record_error();
-        metrics.record_dial_result(Duration::from_millis(9), true);
-        metrics.record_dial_result(Duration::from_millis(3), false);
-
-        let snapshot = metrics.snapshot();
-        assert_eq!(snapshot.sent_bytes_total, 64);
-        assert_eq!(snapshot.recv_bytes_total, 32);
-        assert_eq!(snapshot.outgoing_requests_total, 1);
-        assert_eq!(snapshot.incoming_requests_total, 1);
-        assert_eq!(snapshot.errors_total, 1);
-        assert_eq!(snapshot.dial_errors_total, 1);
-        assert_eq!(snapshot.dial_avg_time_nanos, 6_000_000);
-        assert!(snapshot.last_dial_unix_millis > 0);
-
-        metrics.reset_for_test();
-    }
-}

crates/common/src/lib.rs

@@ -13,12 +13,9 @@
 // limitations under the License.
 
 pub mod bucket_stats;
-pub mod capacity_scope;
 // pub mod error;
-pub mod data_usage;
 pub mod globals;
 pub mod heal_channel;
-pub mod internode_metrics;
 pub mod last_minute;
 pub mod metrics;
 mod readiness;

crates/common/src/metrics.rs

@@ -14,7 +14,6 @@
 
 use crate::last_minute::{AccElem, LastMinuteLatency};
 use chrono::{DateTime, Utc};
-use rustfs_madmin::metrics::{ScannerMetrics as M_ScannerMetrics, TimedAction};
 use serde::{Deserialize, Serialize};
 use std::{
     collections::HashMap,
@@ -351,6 +350,32 @@ pub struct CurrentCycle {
     pub started: DateTime<Utc>,
 }
 
+#[derive(Clone, Debug, Default, Serialize, Deserialize, PartialEq, Eq)]
+pub struct ScannerTimedAction {
+    pub count: u64,
+    pub acc_time: u64,
+    pub bytes: u64,
+}
+
+#[derive(Clone, Debug, Default, Serialize, Deserialize)]
+pub struct ScannerLastMinute {
+    pub actions: HashMap<String, ScannerTimedAction>,
+    pub ilm: HashMap<String, ScannerTimedAction>,
+}
+
+#[derive(Clone, Debug, Default, Serialize, Deserialize)]
+pub struct ScannerMetricsReport {
+    pub collected_at: DateTime<Utc>,
+    pub current_cycle: u64,
+    pub current_started: DateTime<Utc>,
+    pub cycles_completed_at: Vec<DateTime<Utc>>,
+    pub ongoing_buckets: usize,
+    pub life_time_ops: HashMap<String, u64>,
+    pub life_time_ilm: HashMap<String, u64>,
+    pub last_minute: ScannerLastMinute,
+    pub active_paths: Vec<String>,
+}
+
 impl CurrentCycle {
     pub fn unmarshal(&mut self, buf: &[u8]) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
         *self = rmp_serde::from_slice(buf)?;
@@ -572,8 +597,8 @@ impl Metrics {
     }
 
     /// Build a full metrics report snapshot.
-    pub async fn report(&self) -> M_ScannerMetrics {
-        let mut m = M_ScannerMetrics::default();
+    pub async fn report(&self) -> ScannerMetricsReport {
+        let mut m = ScannerMetricsReport::default();
 
         if let Some(cycle) = self.get_cycle().await {
             m.current_cycle = cycle.current;
@@ -606,7 +631,7 @@ impl Metrics {
             {
                 m.last_minute.actions.insert(
                     metric.as_str().to_string(),
-                    TimedAction {
+                    ScannerTimedAction {
                         count: last_min.n,
                         acc_time: last_min.total,
                         bytes: last_min.size,
@@ -633,7 +658,7 @@ impl Metrics {
             {
                 m.last_minute.ilm.insert(
                     action.as_str().to_string(),
-                    TimedAction {
+                    ScannerTimedAction {
                         count: last_min.n,
                         acc_time: last_min.total,
                         bytes: last_min.size,

crates/concurrency/src/backpressure.rs

@@ -14,15 +14,17 @@
 
 //! Backpressure management
 
-use rustfs_io_core::{BackpressureMonitor as CoreBackpressureMonitor, BackpressureState};
+use rustfs_io_core::{
+    BackpressureConfig as CoreBackpressureConfig, BackpressureMonitor as CoreBackpressureMonitor, BackpressureState,
+};
 use rustfs_io_metrics::backpressure_metrics;
 use std::sync::Arc;
 use std::time::Instant;
 use tokio::io::{DuplexStream, duplex};
 
-/// Backpressure configuration
-#[derive(Debug, Clone)]
-pub struct BackpressureConfig {
+/// Facade policy for duplex-pipe watermark backpressure.
+#[derive(Debug, Clone, Copy)]
+pub struct PipeBackpressurePolicy {
     /// Buffer size in bytes
     pub buffer_size: usize,
     /// High watermark percentage
@@ -31,7 +33,7 @@ pub struct BackpressureConfig {
     pub low_watermark: u32,
 }
 
-impl Default for BackpressureConfig {
+impl Default for PipeBackpressurePolicy {
     fn default() -> Self {
         Self {
             buffer_size: 4 * 1024 * 1024, // 4MB
@@ -41,7 +43,7 @@ impl Default for BackpressureConfig {
     }
 }
 
-impl BackpressureConfig {
+impl PipeBackpressurePolicy {
     /// Calculate high watermark threshold in bytes
     pub fn high_watermark_bytes(&self) -> usize {
         (self.buffer_size as u64 * self.high_watermark as u64 / 100) as usize
@@ -51,42 +53,59 @@ impl BackpressureConfig {
     pub fn low_watermark_bytes(&self) -> usize {
         (self.buffer_size as u64 * self.low_watermark as u64 / 100) as usize
     }
+
+    /// Convert the facade policy into the reusable io-core admission-pressure config.
+    ///
+    /// The concurrency layer still owns duplex buffer sizing, but the shared
+    /// overload/admission primitive lives in `io-core`.
+    pub fn to_core_config(&self) -> CoreBackpressureConfig {
+        CoreBackpressureConfig {
+            max_concurrent: 32,
+            high_water_mark: self.high_watermark as f64 / 100.0,
+            low_water_mark: self.low_watermark as f64 / 100.0,
+            cooldown: std::time::Duration::from_millis(100),
+            enabled: true,
+        }
+    }
 }
 
 /// Backpressure manager
 pub struct BackpressureManager {
-    config: BackpressureConfig,
+    config: PipeBackpressurePolicy,
+    core_config: CoreBackpressureConfig,
     monitor: Arc<CoreBackpressureMonitor>,
 }
 
 impl BackpressureManager {
     /// Create a new backpressure manager
     pub fn new(buffer_size: usize, high_watermark: u32, low_watermark: u32) -> Self {
-        let config = BackpressureConfig {
+        Self::from_policy(PipeBackpressurePolicy {
             buffer_size,
             high_watermark,
             low_watermark,
-        };
-
-        let core_config = rustfs_io_core::BackpressureConfig {
-            max_concurrent: 32,
-            high_water_mark: high_watermark as f64 / 100.0,
-            low_water_mark: low_watermark as f64 / 100.0,
-            cooldown: std::time::Duration::from_millis(100),
-            enabled: true,
-        };
+        })
+    }
 
+    /// Create a new backpressure manager from the facade policy type.
+    pub fn from_policy(config: PipeBackpressurePolicy) -> Self {
+        let core_config = config.to_core_config();
         Self {
             config,
+            core_config: core_config.clone(),
             monitor: Arc::new(CoreBackpressureMonitor::new(core_config)),
         }
     }
 
     /// Get the configuration
-    pub fn config(&self) -> &BackpressureConfig {
+    pub fn config(&self) -> &PipeBackpressurePolicy {
         &self.config
     }
 
+    /// Get the derived io-core admission-pressure configuration.
+    pub fn core_config(&self) -> &CoreBackpressureConfig {
+        &self.core_config
+    }
+
     /// Get the monitor
     pub fn monitor(&self) -> Arc<CoreBackpressureMonitor> {
         self.monitor.clone()
@@ -94,7 +113,7 @@ impl BackpressureManager {
 
     /// Create a backpressure pipe
     pub fn create_pipe(&self) -> BackpressurePipe {
-        BackpressurePipe::new(self.config.clone(), self.monitor.clone())
+        BackpressurePipe::new(self.config, self.monitor.clone())
     }
 
     /// Get current state
@@ -112,13 +131,24 @@ impl BackpressureManager {
 pub struct BackpressurePipe {
     reader: DuplexStream,
     writer: DuplexStream,
-    config: BackpressureConfig,
+    config: PipeBackpressurePolicy,
     monitor: Arc<CoreBackpressureMonitor>,
     created_at: Instant,
 }
 
+/// Shared pipe metadata snapshot for facade-level backpressure pipes.
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub struct BackpressurePipeMeta {
+    /// Configured duplex buffer capacity in bytes.
+    pub buffer_capacity: usize,
+    /// Current backpressure state reported by the shared core monitor.
+    pub state: BackpressureState,
+    /// Age of the pipe since creation.
+    pub age: std::time::Duration,
+}
+
 impl BackpressurePipe {
-    fn new(config: BackpressureConfig, monitor: Arc<CoreBackpressureMonitor>) -> Self {
+    fn new(config: PipeBackpressurePolicy, monitor: Arc<CoreBackpressureMonitor>) -> Self {
         let (reader, writer) = duplex(config.buffer_size);
 
         Self {
@@ -146,7 +176,7 @@ impl BackpressurePipe {
     }
 
     /// Get the configuration
-    pub fn config(&self) -> &BackpressureConfig {
+    pub fn config(&self) -> &PipeBackpressurePolicy {
         &self.config
     }
 
@@ -160,6 +190,15 @@ impl BackpressurePipe {
         self.created_at.elapsed()
     }
 
+    /// Get a compact metadata snapshot for the pipe.
+    pub fn meta(&self) -> BackpressurePipeMeta {
+        BackpressurePipeMeta {
+            buffer_capacity: self.config.buffer_size,
+            state: self.state(),
+            age: self.age(),
+        }
+    }
+
     /// Check if should apply backpressure
     pub fn should_apply_backpressure(&self) -> bool {
         let should = self.monitor.should_apply_backpressure();
@@ -170,45 +209,26 @@ impl BackpressurePipe {
     }
 }
 
-/// Backpressure event
-#[allow(dead_code)]
-#[derive(Debug, Clone)]
-pub struct BackpressureEvent {
-    /// Event timestamp
-    pub timestamp: Instant,
-    /// Event type
-    pub event_type: BackpressureEventType,
-    /// Buffer usage
-    pub buffer_usage: usize,
-    /// Buffer capacity
-    pub buffer_capacity: usize,
-}
-
-/// Backpressure event type
-#[allow(dead_code)]
-#[derive(Debug, Clone, Copy)]
-pub enum BackpressureEventType {
-    /// High watermark reached
-    HighWatermarkReached,
-    /// High watermark exited
-    HighWatermarkExited,
-    /// Backpressure applied
-    BackpressureApplied,
-    /// Backpressure released
-    BackpressureReleased,
-}
-
 #[cfg(test)]
 mod tests {
     use super::*;
 
     #[test]
     fn test_backpressure_config() {
-        let config = BackpressureConfig::default();
+        let config = PipeBackpressurePolicy::default();
         assert_eq!(config.buffer_size, 4 * 1024 * 1024);
         assert!(config.high_watermark > config.low_watermark);
     }
 
+    #[test]
+    fn test_backpressure_policy_to_core_config() {
+        let policy = PipeBackpressurePolicy::default();
+        let core = policy.to_core_config();
+        assert_eq!(core.high_water_mark, policy.high_watermark as f64 / 100.0);
+        assert_eq!(core.low_water_mark, policy.low_watermark as f64 / 100.0);
+        assert!(core.enabled);
+    }
+
     #[test]
     fn test_backpressure_manager() {
         let manager = BackpressureManager::new(1024, 80, 50);
@@ -220,5 +240,6 @@ mod tests {
         let manager = BackpressureManager::new(1024, 80, 50);
         let pipe = manager.create_pipe();
         assert_eq!(pipe.state(), BackpressureState::Normal);
+        assert_eq!(pipe.meta().buffer_capacity, 1024);
     }
 }

crates/concurrency/src/config.rs

@@ -14,6 +14,10 @@
 
 //! Configuration for concurrency management
 
+use crate::{
+    backpressure::PipeBackpressurePolicy, deadlock::DeadlockMonitorPolicy, scheduler::SchedulerPolicy,
+    timeout::TimeoutManagerPolicy,
+};
 use std::time::Duration;
 
 /// Feature flags for concurrency modules
@@ -72,84 +76,39 @@ impl ConcurrencyFeatures {
     }
 }
 
+/// Facade policy for lock manager behavior.
+#[derive(Debug, Clone, Copy)]
+pub struct LockManagerPolicy {
+    /// Enable lock optimization.
+    pub enabled: bool,
+    /// Lock acquisition timeout.
+    pub acquire_timeout: Duration,
+}
+
+impl Default for LockManagerPolicy {
+    fn default() -> Self {
+        Self {
+            enabled: true,
+            acquire_timeout: Duration::from_secs(5),
+        }
+    }
+}
+
 /// Main configuration for concurrency management
-#[derive(Debug, Clone)]
+#[derive(Debug, Clone, Default)]
 pub struct ConcurrencyConfig {
     /// Feature flags
     pub features: ConcurrencyFeatures,
-
-    // Timeout configuration
-    /// Default timeout duration
-    pub default_timeout: Duration,
-    /// Maximum timeout duration
-    pub max_timeout: Duration,
-    /// Enable dynamic timeout
-    pub enable_dynamic_timeout: bool,
-
-    // Lock configuration
-    /// Enable lock optimization
-    pub enable_lock_optimization: bool,
-    /// Lock acquisition timeout
-    pub lock_acquire_timeout: Duration,
-
-    // Deadlock configuration
-    /// Enable deadlock detection
-    pub enable_deadlock_detection: bool,
-    /// Deadlock check interval
-    pub deadlock_check_interval: Duration,
-    /// Hang threshold
-    pub hang_threshold: Duration,
-
-    // Backpressure configuration
-    /// Buffer size for backpressure
-    pub backpressure_buffer_size: usize,
-    /// High watermark percentage
-    pub high_watermark: u32,
-    /// Low watermark percentage
-    pub low_watermark: u32,
-
-    // Scheduler configuration
-    /// Base buffer size for I/O
-    pub io_buffer_size: usize,
-    /// Maximum buffer size
-    pub max_buffer_size: usize,
-    /// High priority size threshold
-    pub high_priority_threshold: usize,
-    /// Low priority size threshold
-    pub low_priority_threshold: usize,
-}
-
-impl Default for ConcurrencyConfig {
-    fn default() -> Self {
-        Self {
-            features: ConcurrencyFeatures::default(),
-
-            // Timeout defaults
-            default_timeout: Duration::from_secs(30),
-            max_timeout: Duration::from_secs(300),
-            enable_dynamic_timeout: true,
-
-            // Lock defaults
-            enable_lock_optimization: true,
-            lock_acquire_timeout: Duration::from_secs(5),
-
-            // Deadlock defaults
-            enable_deadlock_detection: false,
-            deadlock_check_interval: Duration::from_secs(10),
-            hang_threshold: Duration::from_secs(60),
-
-            // Backpressure defaults
-            backpressure_buffer_size: 4 * 1024 * 1024, // 4MB
-            high_watermark: 80,
-            low_watermark: 50,
-
-            // Scheduler defaults
-            io_buffer_size: 64 * 1024,                // 64KB
-            max_buffer_size: 4 * 1024 * 1024,         // 4MB
-            high_priority_threshold: 1024 * 1024,     // 1MB
-            low_priority_threshold: 10 * 1024 * 1024, // 10MB
-        }
-    }
+    /// Timeout facade policy.
+    pub timeout_policy: TimeoutManagerPolicy,
+    /// Lock facade policy.
+    pub lock_policy: LockManagerPolicy,
+    /// Deadlock facade policy.
+    pub deadlock_policy: DeadlockMonitorPolicy,
+    /// Backpressure facade policy.
+    pub backpressure_policy: PipeBackpressurePolicy,
+    /// Scheduler facade policy.
+    pub scheduler_policy: SchedulerPolicy,
 }
 
 impl ConcurrencyConfig {
@@ -161,25 +120,25 @@ impl ConcurrencyConfig {
         if let Ok(val) = std::env::var("RUSTFS_TIMEOUT_DEFAULT")
             && let Ok(secs) = val.parse::<u64>()
         {
-            config.default_timeout = Duration::from_secs(secs);
+            config.timeout_policy.default_timeout = Duration::from_secs(secs);
         }
 
         if let Ok(val) = std::env::var("RUSTFS_TIMEOUT_MAX")
             && let Ok(secs) = val.parse::<u64>()
         {
-            config.max_timeout = Duration::from_secs(secs);
+            config.timeout_policy.max_timeout = Duration::from_secs(secs);
         }
 
         if let Ok(val) = std::env::var("RUSTFS_BACKPRESSURE_BUFFER_SIZE")
             && let Ok(size) = val.parse::<usize>()
         {
-            config.backpressure_buffer_size = size;
+            config.backpressure_policy.buffer_size = size;
         }
 
         if let Ok(val) = std::env::var("RUSTFS_IO_BUFFER_SIZE")
             && let Ok(size) = val.parse::<usize>()
         {
-            config.io_buffer_size = size;
+            config.scheduler_policy.base_buffer_size = size;
         }
 
         config
@@ -187,18 +146,25 @@ impl ConcurrencyConfig {
 
     /// Validate configuration
     pub fn validate(&self) -> Result<(), ConfigError> {
-        if self.default_timeout > self.max_timeout {
+        if self.timeout_policy.default_timeout > self.timeout_policy.max_timeout {
             return Err(ConfigError::InvalidTimeout("default_timeout cannot exceed max_timeout".to_string()));
         }
+        if self.timeout_policy.min_timeout > self.timeout_policy.max_timeout {
+            return Err(ConfigError::InvalidTimeout("min_timeout cannot exceed max_timeout".to_string()));
+        }
 
-        if self.high_watermark <= self.low_watermark || self.high_watermark > 100 {
+        if self.backpressure_policy.high_watermark <= self.backpressure_policy.low_watermark
+            || self.backpressure_policy.high_watermark > 100
+        {
             return Err(ConfigError::InvalidBackpressure(
                 "high_watermark must be > low_watermark and <= 100".to_string(),
             ));
         }
 
-        if self.io_buffer_size > self.max_buffer_size {
-            return Err(ConfigError::InvalidScheduler("io_buffer_size cannot exceed max_buffer_size".to_string()));
+        if self.scheduler_policy.base_buffer_size > self.scheduler_policy.max_buffer_size {
+            return Err(ConfigError::InvalidScheduler(
+                "base_buffer_size cannot exceed max_buffer_size".to_string(),
+            ));
         }
 
         Ok(())
@@ -235,8 +201,12 @@ mod tests {
     #[test]
     fn test_invalid_timeout() {
         let config = ConcurrencyConfig {
-            default_timeout: Duration::from_secs(100),
-            max_timeout: Duration::from_secs(50),
+            timeout_policy: TimeoutManagerPolicy {
+                default_timeout: Duration::from_secs(100),
+                max_timeout: Duration::from_secs(50),
+                enable_dynamic: true,
+                ..Default::default()
+            },
             ..Default::default()
         };
         assert!(
@@ -245,6 +215,22 @@ mod tests {
         );
     }
 
+    #[test]
+    fn test_invalid_min_timeout() {
+        let config = ConcurrencyConfig {
+            timeout_policy: TimeoutManagerPolicy {
+                min_timeout: Duration::from_secs(100),
+                max_timeout: Duration::from_secs(50),
+                ..Default::default()
+            },
+            ..Default::default()
+        };
+        assert!(
+            config.validate().is_err(),
+            "validate() should return an error when min_timeout > max_timeout"
+        );
+    }
+
     #[test]
     fn test_features() {
         let features = ConcurrencyFeatures::all();

crates/concurrency/src/deadlock.rs

@@ -14,15 +14,15 @@
 
 //! Deadlock detection management
 
-use rustfs_io_core::{DeadlockDetector as CoreDeadlockDetector, LockType};
+use rustfs_io_core::{DeadlockDetector as CoreDeadlockDetector, DeadlockDetectorConfig as CoreDeadlockConfig, LockType};
 use rustfs_io_metrics::deadlock_metrics;
 use std::collections::HashMap;
 use std::sync::Arc;
 use std::time::{Duration, Instant};
 
-/// Deadlock configuration
-#[derive(Debug, Clone)]
-pub struct DeadlockConfig {
+/// Facade policy for the concurrency-layer deadlock monitor.
+#[derive(Debug, Clone, Copy)]
+pub struct DeadlockMonitorPolicy {
     /// Enable deadlock detection
     pub enabled: bool,
     /// Check interval
@@ -31,7 +31,7 @@ pub struct DeadlockConfig {
     pub hang_threshold: Duration,
 }
 
-impl Default for DeadlockConfig {
+impl Default for DeadlockMonitorPolicy {
     fn default() -> Self {
         Self {
             enabled: false,
@@ -41,9 +41,20 @@ impl Default for DeadlockConfig {
     }
 }
 
+impl DeadlockMonitorPolicy {
+    /// Convert the facade policy into the reusable io-core deadlock config.
+    pub fn to_core_config(&self) -> CoreDeadlockConfig {
+        CoreDeadlockConfig {
+            enabled: self.enabled,
+            detection_interval: self.check_interval,
+            max_hold_time: self.hang_threshold,
+        }
+    }
+}
+
 /// Deadlock manager
 pub struct DeadlockManager {
-    config: DeadlockConfig,
+    config: DeadlockMonitorPolicy,
     detector: Arc<CoreDeadlockDetector>,
     running: Arc<tokio::sync::Mutex<bool>>,
 }
@@ -51,18 +62,16 @@ pub struct DeadlockManager {
 impl DeadlockManager {
     /// Create a new deadlock manager
     pub fn new(enabled: bool, check_interval: Duration, hang_threshold: Duration) -> Self {
-        let config = DeadlockConfig {
+        Self::from_policy(DeadlockMonitorPolicy {
             enabled,
             check_interval,
             hang_threshold,
-        };
-
-        let core_config = rustfs_io_core::DeadlockDetectorConfig {
-            enabled,
-            detection_interval: check_interval,
-            max_hold_time: hang_threshold,
-        };
+        })
+    }
 
+    /// Create a new deadlock manager from the facade policy type.
+    pub fn from_policy(config: DeadlockMonitorPolicy) -> Self {
+        let core_config = config.to_core_config();
         Self {
             config,
             detector: Arc::new(CoreDeadlockDetector::new(core_config)),
@@ -71,7 +80,7 @@ impl DeadlockManager {
     }
 
     /// Get the configuration
-    pub fn config(&self) -> &DeadlockConfig {
+    pub fn config(&self) -> &DeadlockMonitorPolicy {
         &self.config
     }
 
@@ -129,7 +138,11 @@ impl DeadlockManager {
     }
 }
 
-/// Request tracker for tracking resources
+/// Lightweight compatibility wrapper for request-scoped deadlock bookkeeping.
+///
+/// This type intentionally stays minimal in the concurrency layer. Rich
+/// request-level lock/resource diagnostics belong to
+/// `rustfs::storage::deadlock_detector::RequestResourceTracker`.
 pub struct RequestTracker {
     request_id: String,
     description: String,
@@ -174,6 +187,11 @@ impl RequestTracker {
         deadlock_metrics::record_lock_acquisition("read");
     }
 
+    /// Return a read-only view of tracked resource names.
+    pub fn resources(&self) -> &HashMap<String, Vec<String>> {
+        &self.resources
+    }
+
     /// Record a lock release
     pub fn record_lock_release(&mut self, lock_id: u64) {
         self.detector.record_release(lock_id);
@@ -196,12 +214,24 @@ mod tests {
         assert!(!manager.config().enabled);
     }
 
+    #[test]
+    fn test_deadlock_policy_to_core_config() {
+        let policy = DeadlockMonitorPolicy::default();
+        let core = policy.to_core_config();
+        assert_eq!(core.enabled, policy.enabled);
+        assert_eq!(core.detection_interval, policy.check_interval);
+        assert_eq!(core.max_hold_time, policy.hang_threshold);
+    }
+
     #[tokio::test]
     async fn test_request_tracker() {
         let manager = DeadlockManager::new(true, Duration::from_secs(10), Duration::from_secs(60));
-        let tracker = manager.track_request("req-1".to_string(), "test request".to_string());
+        let mut tracker = manager.track_request("req-1".to_string(), "test request".to_string());
+        let lock_id = manager.register_lock(LockType::Mutex);
+        tracker.record_lock_acquire(lock_id, "bucket/key".to_string());
 
         assert_eq!(tracker.request_id(), "req-1");
         assert_eq!(tracker.description(), "test request");
+        assert_eq!(tracker.resources().get("locks").map(Vec::len), Some(1));
     }
 }

crates/concurrency/src/lib.rs

@@ -124,21 +124,23 @@ mod backpressure;
 #[cfg(feature = "scheduler")]
 mod scheduler;
 
+pub mod workers;
+
 // Public module exports with feature gates
 #[cfg(feature = "timeout")]
-pub use timeout::{TimeoutConfig, TimeoutGuard, TimeoutManager};
+pub use timeout::{TimeoutGuard, TimeoutManager, TimeoutManagerPolicy};
 
 #[cfg(feature = "lock")]
 pub use lock::{LockConfig, LockManager, LockScopeGuard, OptimizedLockGuard};
 
 #[cfg(feature = "deadlock")]
-pub use deadlock::{DeadlockConfig, DeadlockManager, RequestTracker};
+pub use deadlock::{DeadlockManager, DeadlockMonitorPolicy, RequestTracker};
 
 #[cfg(feature = "backpressure")]
-pub use backpressure::{BackpressureConfig, BackpressureManager, BackpressurePipe};
+pub use backpressure::{BackpressureManager, BackpressurePipe, PipeBackpressurePolicy};
 
 #[cfg(feature = "scheduler")]
-pub use scheduler::{IoStrategy, SchedulerConfig, SchedulerManager};
+pub use scheduler::{IoStrategy, SchedulerManager, SchedulerPolicy};
 
 // Configuration
 mod config;
@@ -153,19 +155,19 @@ pub mod prelude {
     //! Prelude module for convenient imports
 
     #[cfg(feature = "timeout")]
-    pub use crate::timeout::{TimeoutConfig, TimeoutGuard, TimeoutManager};
+    pub use crate::timeout::{TimeoutGuard, TimeoutManager, TimeoutManagerPolicy};
 
     #[cfg(feature = "lock")]
     pub use crate::lock::{LockConfig, LockManager, LockScopeGuard, OptimizedLockGuard};
 
     #[cfg(feature = "deadlock")]
-    pub use crate::deadlock::{DeadlockConfig, DeadlockManager, RequestTracker};
+    pub use crate::deadlock::{DeadlockManager, DeadlockMonitorPolicy, RequestTracker};
 
     #[cfg(feature = "backpressure")]
-    pub use crate::backpressure::{BackpressureConfig, BackpressureManager, BackpressurePipe};
+    pub use crate::backpressure::{BackpressureManager, BackpressurePipe, PipeBackpressurePolicy};
 
     #[cfg(feature = "scheduler")]
-    pub use crate::scheduler::{IoStrategy, SchedulerConfig, SchedulerManager};
+    pub use crate::scheduler::{IoStrategy, SchedulerManager, SchedulerPolicy};
 
     pub use crate::{ConcurrencyConfig, ConcurrencyFeatures, ConcurrencyManager};
 }

crates/concurrency/src/manager.rs

@@ -85,39 +85,22 @@ impl ConcurrencyManager {
 
         Self {
             #[cfg(feature = "timeout")]
-            timeout: Arc::new(crate::timeout::TimeoutManager::new(
-                config.default_timeout,
-                config.max_timeout,
-                config.enable_dynamic_timeout,
-            )),
+            timeout: Arc::new(crate::timeout::TimeoutManager::from_policy(config.timeout_policy)),
 
             #[cfg(feature = "lock")]
             lock: Arc::new(crate::lock::LockManager::new(
-                config.enable_lock_optimization,
-                config.lock_acquire_timeout,
+                config.lock_policy.enabled,
+                config.lock_policy.acquire_timeout,
             )),
 
             #[cfg(feature = "deadlock")]
-            deadlock: Arc::new(crate::deadlock::DeadlockManager::new(
-                config.enable_deadlock_detection,
-                config.deadlock_check_interval,
-                config.hang_threshold,
-            )),
+            deadlock: Arc::new(crate::deadlock::DeadlockManager::from_policy(config.deadlock_policy)),
 
             #[cfg(feature = "backpressure")]
-            backpressure: Arc::new(crate::backpressure::BackpressureManager::new(
-                config.backpressure_buffer_size,
-                config.high_watermark,
-                config.low_watermark,
-            )),
+            backpressure: Arc::new(crate::backpressure::BackpressureManager::from_policy(config.backpressure_policy)),
 
             #[cfg(feature = "scheduler")]
-            scheduler: Arc::new(crate::scheduler::SchedulerManager::new(
-                config.io_buffer_size,
-                config.max_buffer_size,
-                config.high_priority_threshold,
-                config.low_priority_threshold,
-            )),
+            scheduler: Arc::new(crate::scheduler::SchedulerManager::from_policy(config.scheduler_policy)),
 
             config,
         }
@@ -244,7 +227,7 @@ impl ConcurrencyManager {
     pub async fn start(&self) {
         #[cfg(feature = "deadlock")]
         {
-            if self.config.enable_deadlock_detection {
+            if self.config.deadlock_policy.enabled {
                 self.deadlock.start().await;
             }
         }

crates/concurrency/src/scheduler.rs

@@ -22,9 +22,9 @@ use rustfs_io_metrics::io_metrics;
 use std::sync::Arc;
 use std::time::Duration;
 
-/// Scheduler configuration
-#[derive(Debug, Clone)]
-pub struct SchedulerConfig {
+/// Facade policy for the concurrency-layer scheduler manager.
+#[derive(Debug, Clone, Copy)]
+pub struct SchedulerPolicy {
     /// Base buffer size
     pub base_buffer_size: usize,
     /// Maximum buffer size
@@ -35,7 +35,7 @@ pub struct SchedulerConfig {
     pub low_priority_threshold: usize,
 }
 
-impl Default for SchedulerConfig {
+impl Default for SchedulerPolicy {
     fn default() -> Self {
         Self {
             base_buffer_size: 64 * 1024,              // 64KB
@@ -46,9 +46,23 @@ impl Default for SchedulerConfig {
     }
 }
 
+impl SchedulerPolicy {
+    /// Convert facade policy to io-core scheduler config.
+    pub fn to_core_config(&self) -> rustfs_io_core::IoSchedulerConfig {
+        rustfs_io_core::IoSchedulerConfig {
+            base_buffer_size: self.base_buffer_size,
+            max_buffer_size: self.max_buffer_size,
+            high_priority_size_threshold: self.high_priority_threshold,
+            low_priority_size_threshold: self.low_priority_threshold,
+            ..rustfs_io_core::IoSchedulerConfig::default()
+        }
+    }
+}
+
 /// Scheduler manager
 pub struct SchedulerManager {
-    config: SchedulerConfig,
+    config: SchedulerPolicy,
+    core_config: rustfs_io_core::IoSchedulerConfig,
     scheduler: Arc<CoreIoScheduler>,
 }
 
@@ -60,26 +74,35 @@ impl SchedulerManager {
         high_priority_threshold: usize,
         low_priority_threshold: usize,
     ) -> Self {
-        let config = SchedulerConfig {
+        Self::from_policy(SchedulerPolicy {
             base_buffer_size,
             max_buffer_size,
             high_priority_threshold,
             low_priority_threshold,
-        };
+        })
+    }
 
-        let core_config = rustfs_io_core::IoSchedulerConfig::default();
+    /// Create a scheduler manager from facade policy.
+    pub fn from_policy(config: SchedulerPolicy) -> Self {
+        let core_config = config.to_core_config();
 
         Self {
             config,
+            core_config: core_config.clone(),
             scheduler: Arc::new(CoreIoScheduler::new(core_config)),
         }
     }
 
     /// Get the configuration
-    pub fn config(&self) -> &SchedulerConfig {
+    pub fn config(&self) -> &SchedulerPolicy {
         &self.config
     }
 
+    /// Get the derived io-core scheduler config.
+    pub fn core_config(&self) -> &rustfs_io_core::IoSchedulerConfig {
+        &self.core_config
+    }
+
     /// Get the scheduler
     pub fn scheduler(&self) -> Arc<CoreIoScheduler> {
         self.scheduler.clone()
@@ -87,7 +110,7 @@ impl SchedulerManager {
 
     /// Create an I/O strategy
     pub fn create_strategy(&self) -> IoStrategy {
-        IoStrategy::new(self.config.clone(), self.scheduler.clone())
+        IoStrategy::new(self.config, self.scheduler.clone())
     }
 
     /// Calculate buffer size
@@ -111,12 +134,12 @@ impl SchedulerManager {
 
 /// I/O strategy
 pub struct IoStrategy {
-    config: SchedulerConfig,
+    config: SchedulerPolicy,
     scheduler: Arc<CoreIoScheduler>,
 }
 
 impl IoStrategy {
-    fn new(config: SchedulerConfig, scheduler: Arc<CoreIoScheduler>) -> Self {
+    fn new(config: SchedulerPolicy, scheduler: Arc<CoreIoScheduler>) -> Self {
         Self { config, scheduler }
     }
 
@@ -191,7 +214,7 @@ impl IoStrategy {
     }
 
     /// Get the configuration
-    pub fn config(&self) -> &SchedulerConfig {
+    pub fn config(&self) -> &SchedulerPolicy {
         &self.config
     }
 }
@@ -202,10 +225,20 @@ mod tests {
 
     #[test]
     fn test_scheduler_config() {
-        let config = SchedulerConfig::default();
+        let config = SchedulerPolicy::default();
         assert!(config.base_buffer_size < config.max_buffer_size);
     }
 
+    #[test]
+    fn test_scheduler_policy_to_core_config() {
+        let policy = SchedulerPolicy::default();
+        let core = policy.to_core_config();
+        assert_eq!(core.base_buffer_size, policy.base_buffer_size);
+        assert_eq!(core.max_buffer_size, policy.max_buffer_size);
+        assert_eq!(core.high_priority_size_threshold, policy.high_priority_threshold);
+        assert_eq!(core.low_priority_size_threshold, policy.low_priority_threshold);
+    }
+
     #[test]
     fn test_scheduler_manager() {
         let manager = SchedulerManager::new(1024, 4096, 512, 2048);

crates/concurrency/src/timeout.rs

@@ -14,60 +14,101 @@
 
 //! Timeout management for operations
 
-use rustfs_io_core::{TimeoutError, calculate_adaptive_timeout};
+use rustfs_io_core::{TimeoutConfig as CoreTimeoutConfig, TimeoutError, calculate_adaptive_timeout};
 use std::time::{Duration, Instant};
 use tokio_util::sync::CancellationToken;
 
-/// Timeout configuration
-#[derive(Debug, Clone)]
-pub struct TimeoutConfig {
+/// Facade policy for the concurrency-layer timeout manager.
+#[derive(Debug, Clone, Copy)]
+pub struct TimeoutManagerPolicy {
     /// Default timeout duration
     pub default_timeout: Duration,
     /// Maximum timeout duration
     pub max_timeout: Duration,
+    /// Minimum timeout floor (prevents dynamic calculation from going too low).
+    pub min_timeout: Duration,
     /// Enable dynamic timeout calculation
     pub enable_dynamic: bool,
 }
 
-impl Default for TimeoutConfig {
+impl Default for TimeoutManagerPolicy {
     fn default() -> Self {
         Self {
             default_timeout: Duration::from_secs(30),
             max_timeout: Duration::from_secs(300),
+            min_timeout: Duration::from_secs(5),
             enable_dynamic: true,
         }
     }
 }
 
+impl TimeoutManagerPolicy {
+    /// Convert the facade policy into the reusable io-core timeout configuration.
+    ///
+    /// This keeps the concurrency layer explicitly wired to the shared core
+    /// timeout primitives without changing the facade's public behavior.
+    pub fn to_core_config(&self) -> CoreTimeoutConfig {
+        CoreTimeoutConfig {
+            base_timeout: self.default_timeout,
+            timeout_per_mb: Duration::ZERO,
+            max_timeout: self.max_timeout,
+            min_timeout: self.min_timeout,
+            get_object_timeout: self.default_timeout,
+            put_object_timeout: self.max_timeout,
+            list_objects_timeout: self.default_timeout,
+            enable_dynamic_timeout: self.enable_dynamic,
+        }
+    }
+}
+
 /// Timeout manager
 pub struct TimeoutManager {
-    config: TimeoutConfig,
+    config: TimeoutManagerPolicy,
+    core_config: CoreTimeoutConfig,
 }
 
 impl TimeoutManager {
     /// Create a new timeout manager
     pub fn new(default_timeout: Duration, max_timeout: Duration, enable_dynamic: bool) -> Self {
-        Self {
-            config: TimeoutConfig {
-                default_timeout,
-                max_timeout,
-                enable_dynamic,
-            },
-        }
+        let min_timeout = default_timeout.min(max_timeout);
+        Self::from_policy(TimeoutManagerPolicy {
+            default_timeout,
+            max_timeout,
+            min_timeout,
+            enable_dynamic,
+        })
+    }
+
+    /// Create a new timeout manager from the facade policy type.
+    pub fn from_policy(config: TimeoutManagerPolicy) -> Self {
+        let config = TimeoutManagerPolicy {
+            // Guard clamp(min, max) from panic when callers provide an
+            // out-of-order policy (or very small max_timeout).
+            min_timeout: config.min_timeout.min(config.max_timeout),
+            ..config
+        };
+        let core_config = config.to_core_config();
+        Self { config, core_config }
     }
 
     /// Get the configuration
-    pub fn config(&self) -> &TimeoutConfig {
+    pub fn config(&self) -> &TimeoutManagerPolicy {
         &self.config
     }
 
+    /// Get the derived io-core timeout configuration.
+    pub fn core_config(&self) -> &CoreTimeoutConfig {
+        &self.core_config
+    }
+
     /// Calculate timeout for a given size
     pub fn calculate_timeout(&self, size: u64, _history: &[Duration]) -> Duration {
         if !self.config.enable_dynamic {
             return self.config.default_timeout;
         }
 
-        calculate_adaptive_timeout(self.config.default_timeout, None, 0, size).min(self.config.max_timeout)
+        calculate_adaptive_timeout(self.core_config.base_timeout, None, 0, size)
+            .clamp(self.core_config.min_timeout, self.core_config.max_timeout)
     }
 
     /// Wrap an operation with timeout control
@@ -87,7 +128,7 @@ impl TimeoutManager {
 
     /// Create a timeout guard for manual timeout control
     pub fn create_guard(&self, timeout: Option<Duration>) -> TimeoutGuard {
-        TimeoutGuard::new(timeout.unwrap_or(self.config.default_timeout))
+        TimeoutGuard::new(timeout.unwrap_or(self.core_config.base_timeout))
     }
 }
 
@@ -134,10 +175,41 @@ mod tests {
 
     #[test]
     fn test_timeout_config() {
-        let config = TimeoutConfig::default();
+        let config = TimeoutManagerPolicy::default();
         assert!(config.default_timeout < config.max_timeout);
     }
 
+    #[test]
+    fn test_timeout_policy_to_core_config() {
+        let policy = TimeoutManagerPolicy::default();
+        let core = policy.to_core_config();
+        assert_eq!(core.base_timeout, policy.default_timeout);
+        assert_eq!(core.max_timeout, policy.max_timeout);
+        assert_eq!(core.min_timeout, policy.min_timeout);
+        assert_eq!(core.get_object_timeout, policy.default_timeout);
+        assert!(core.enable_dynamic_timeout);
+    }
+
+    #[test]
+    fn test_timeout_manager_new_sanitizes_min_timeout_with_small_max_timeout() {
+        let manager = TimeoutManager::new(Duration::from_secs(1), Duration::from_secs(1), true);
+        let timeout = manager.calculate_timeout(1024, &[]);
+        assert_eq!(timeout, Duration::from_secs(1));
+    }
+
+    #[test]
+    fn test_timeout_manager_from_policy_sanitizes_min_timeout() {
+        let manager = TimeoutManager::from_policy(TimeoutManagerPolicy {
+            default_timeout: Duration::from_secs(30),
+            max_timeout: Duration::from_secs(1),
+            min_timeout: Duration::from_secs(5),
+            enable_dynamic: true,
+        });
+
+        assert_eq!(manager.config().min_timeout, Duration::from_secs(1));
+        assert_eq!(manager.core_config().min_timeout, Duration::from_secs(1));
+    }
+
     #[tokio::test]
     async fn test_wrap_operation_success() {
         let manager = TimeoutManager::new(Duration::from_secs(5), Duration::from_secs(10), true);

crates/concurrency/src/workers.rs

@@ -12,10 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+//! Worker slot limiter used by long-running background workflows.
+
 use std::sync::Arc;
 use tokio::sync::{Mutex, Notify};
 use tracing::info;
 
+/// Cooperative worker-slot controller for async tasks.
 pub struct Workers {
     available: Mutex<usize>, // Available working slots
     notify: Notify,          // Used to notify waiting tasks
@@ -23,20 +26,20 @@ pub struct Workers {
 }
 
 impl Workers {
-    // Create a Workers object that allows up to n jobs to execute concurrently
-    pub fn new(n: usize) -> Result<Arc<Workers>, &'static str> {
+    /// Create a [`Workers`] object that allows up to `n` jobs to execute concurrently.
+    pub fn new(n: usize) -> Result<Arc<Self>, &'static str> {
         if n == 0 {
             return Err("n must be > 0");
         }
 
-        Ok(Arc::new(Workers {
+        Ok(Arc::new(Self {
             available: Mutex::new(n),
             notify: Notify::new(),
             limit: n,
         }))
     }
 
-    // Give a job a chance to be executed
+    /// Acquire a worker slot, waiting until one becomes available.
     pub async fn take(&self) {
         loop {
             let mut available = self.available.lock().await;
@@ -51,7 +54,7 @@ impl Workers {
         }
     }
 
-    // Release a job's slot
+    /// Release a worker slot.
     pub async fn give(&self) {
         let mut available = self.available.lock().await;
         info!("worker give, {}", *available);
@@ -59,7 +62,7 @@ impl Workers {
         self.notify.notify_one(); // Notify a waiting task
     }
 
-    // Wait for all concurrent jobs to complete
+    /// Wait until all worker slots are released.
     pub async fn wait(&self) {
         loop {
             {
@@ -74,6 +77,7 @@ impl Workers {
         info!("worker wait end");
     }
 
+    /// Return the current number of available worker slots.
     pub async fn available(&self) -> usize {
         *self.available.lock().await
     }

crates/config/README.md

@@ -41,6 +41,7 @@ Examples:
 - `RUSTFS_ADDRESS`
 - `RUSTFS_VOLUMES`
 - `RUSTFS_LICENSE`
+- `RUSTFS_LICENSE_PUBLIC_KEY`
 
 Current guidance:
 - Prefer module-specific names only when they are not top-level product configuration.
@@ -51,6 +52,16 @@ Current guidance:
   - `RUSTFS_ENABLE_HEAL` -> `RUSTFS_HEAL_ENABLED`
   - `RUSTFS_DATA_SCANNER_START_DELAY_SECS` -> `RUSTFS_SCANNER_START_DELAY_SECS`
 
+## License environment variables
+
+- `RUSTFS_LICENSE` contains the signed license token.
+- `RUSTFS_LICENSE_PUBLIC_KEY` contains the RSA public key used to verify signed license tokens.
+
+## CORS environment variables
+
+- `RUSTFS_CORS_ALLOWED_ORIGINS` defaults to empty, so the S3 endpoint emits no generic CORS headers unless configured. Set `*` for wildcard origins without credentials, or a comma-separated allow-list for credentialed explicit origins.
+- `RUSTFS_CONSOLE_CORS_ALLOWED_ORIGINS` defaults to `*` for the console service.
+
 ## Scanner environment aliases
 
 - `RUSTFS_SCANNER_SPEED` (canonical, also accepts `MINIO_SCANNER_SPEED`)
@@ -72,6 +83,28 @@ Legacy compatibility fallback:
 - `RUSTFS_DRIVE_MAX_TIMEOUT_DURATION`
   This legacy variable is treated as a deprecated fallback for the operation-specific drive timeout variables above when a canonical variable is unset.
 
+Drive timeout health-action policy:
+- `RUSTFS_DRIVE_TIMEOUT_HEALTH_ACTION`
+  - `mark_failure` (default): timeout marks failure and may transition drive runtime state.
+  - `ignore_scanner`: timeout does not mark failure for scanner-sensitive operations (`walk_dir`, `read_metadata`, `list_dir`, `disk_info`).
+
+Drive timeout profile preset:
+- `RUSTFS_DRIVE_TIMEOUT_PROFILE`
+  - `default` (default): keep current timeout defaults.
+  - `high_latency`: use 60s default timeout for scanner-sensitive operations when no per-operation timeout override is set (`read_metadata`, `disk_info`, `list_dir`, `walk_dir`, `walk_dir_stall`).
+- Precedence:
+  - Explicit per-operation timeout env (`RUSTFS_DRIVE_*_TIMEOUT_SECS`) takes highest precedence.
+  - Then `RUSTFS_DRIVE_MAX_TIMEOUT_DURATION` legacy fallback.
+  - Then the profile-derived default (`default` or `high_latency`).
+
+## Startup filesystem boundary policy
+
+- `RUSTFS_UNSUPPORTED_FS_POLICY` controls startup behavior when RustFS detects local endpoint filesystems that are outside the supported production boundary.
+  - `warn` (default): log warning and continue startup.
+  - `fail`: abort startup with an error.
+
+RustFS production guidance remains direct-attached local POSIX filesystems. Network-mounted filesystems (for example `nfs`, `cifs`, `smb2`, and `fuse.*`) are treated as unsupported by this startup guard.
+
 ## 📄 License
 
 This project is licensed under the Apache License 2.0 - see the [LICENSE](../../LICENSE) file for details.

crates/config/src/audit/amqp.rs

@@ -0,0 +1,60 @@
+// Copyright 2024 RustFS Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+pub const AUDIT_AMQP_KEYS: &[&str] = &[
+    crate::ENABLE_KEY,
+    crate::AMQP_URL,
+    crate::AMQP_EXCHANGE,
+    crate::AMQP_ROUTING_KEY,
+    crate::AMQP_MANDATORY,
+    crate::AMQP_PERSISTENT,
+    crate::AMQP_USERNAME,
+    crate::AMQP_PASSWORD,
+    crate::AMQP_TLS_CA,
+    crate::AMQP_TLS_CLIENT_CERT,
+    crate::AMQP_TLS_CLIENT_KEY,
+    crate::AMQP_QUEUE_DIR,
+    crate::AMQP_QUEUE_LIMIT,
+    crate::COMMENT_KEY,
+];
+
+pub const ENV_AUDIT_AMQP_ENABLE: &str = "RUSTFS_AUDIT_AMQP_ENABLE";
+pub const ENV_AUDIT_AMQP_URL: &str = "RUSTFS_AUDIT_AMQP_URL";
+pub const ENV_AUDIT_AMQP_EXCHANGE: &str = "RUSTFS_AUDIT_AMQP_EXCHANGE";
+pub const ENV_AUDIT_AMQP_ROUTING_KEY: &str = "RUSTFS_AUDIT_AMQP_ROUTING_KEY";
+pub const ENV_AUDIT_AMQP_MANDATORY: &str = "RUSTFS_AUDIT_AMQP_MANDATORY";
+pub const ENV_AUDIT_AMQP_PERSISTENT: &str = "RUSTFS_AUDIT_AMQP_PERSISTENT";
+pub const ENV_AUDIT_AMQP_USERNAME: &str = "RUSTFS_AUDIT_AMQP_USERNAME";
+pub const ENV_AUDIT_AMQP_PASSWORD: &str = "RUSTFS_AUDIT_AMQP_PASSWORD";
+pub const ENV_AUDIT_AMQP_TLS_CA: &str = "RUSTFS_AUDIT_AMQP_TLS_CA";
+pub const ENV_AUDIT_AMQP_TLS_CLIENT_CERT: &str = "RUSTFS_AUDIT_AMQP_TLS_CLIENT_CERT";
+pub const ENV_AUDIT_AMQP_TLS_CLIENT_KEY: &str = "RUSTFS_AUDIT_AMQP_TLS_CLIENT_KEY";
+pub const ENV_AUDIT_AMQP_QUEUE_DIR: &str = "RUSTFS_AUDIT_AMQP_QUEUE_DIR";
+pub const ENV_AUDIT_AMQP_QUEUE_LIMIT: &str = "RUSTFS_AUDIT_AMQP_QUEUE_LIMIT";
+
+pub const ENV_AUDIT_AMQP_KEYS: &[&str; 13] = &[
+    ENV_AUDIT_AMQP_ENABLE,
+    ENV_AUDIT_AMQP_URL,
+    ENV_AUDIT_AMQP_EXCHANGE,
+    ENV_AUDIT_AMQP_ROUTING_KEY,
+    ENV_AUDIT_AMQP_MANDATORY,
+    ENV_AUDIT_AMQP_PERSISTENT,
+    ENV_AUDIT_AMQP_USERNAME,
+    ENV_AUDIT_AMQP_PASSWORD,
+    ENV_AUDIT_AMQP_TLS_CA,
+    ENV_AUDIT_AMQP_TLS_CLIENT_CERT,
+    ENV_AUDIT_AMQP_TLS_CLIENT_KEY,
+    ENV_AUDIT_AMQP_QUEUE_DIR,
+    ENV_AUDIT_AMQP_QUEUE_LIMIT,
+];

crates/config/src/audit/kafka.rs

@@ -0,0 +1,53 @@
+// Copyright 2024 RustFS Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// Kafka Environment Variables
+pub const ENV_AUDIT_KAFKA_ENABLE: &str = "RUSTFS_AUDIT_KAFKA_ENABLE";
+pub const ENV_AUDIT_KAFKA_BROKERS: &str = "RUSTFS_AUDIT_KAFKA_BROKERS";
+pub const ENV_AUDIT_KAFKA_TOPIC: &str = "RUSTFS_AUDIT_KAFKA_TOPIC";
+pub const ENV_AUDIT_KAFKA_ACKS: &str = "RUSTFS_AUDIT_KAFKA_ACKS";
+pub const ENV_AUDIT_KAFKA_TLS_ENABLE: &str = "RUSTFS_AUDIT_KAFKA_TLS_ENABLE";
+pub const ENV_AUDIT_KAFKA_TLS_CA: &str = "RUSTFS_AUDIT_KAFKA_TLS_CA";
+pub const ENV_AUDIT_KAFKA_TLS_CLIENT_CERT: &str = "RUSTFS_AUDIT_KAFKA_TLS_CLIENT_CERT";
+pub const ENV_AUDIT_KAFKA_TLS_CLIENT_KEY: &str = "RUSTFS_AUDIT_KAFKA_TLS_CLIENT_KEY";
+pub const ENV_AUDIT_KAFKA_QUEUE_DIR: &str = "RUSTFS_AUDIT_KAFKA_QUEUE_DIR";
+pub const ENV_AUDIT_KAFKA_QUEUE_LIMIT: &str = "RUSTFS_AUDIT_KAFKA_QUEUE_LIMIT";
+
+pub const ENV_AUDIT_KAFKA_KEYS: &[&str; 10] = &[
+    ENV_AUDIT_KAFKA_ENABLE,
+    ENV_AUDIT_KAFKA_BROKERS,
+    ENV_AUDIT_KAFKA_TOPIC,
+    ENV_AUDIT_KAFKA_ACKS,
+    ENV_AUDIT_KAFKA_TLS_ENABLE,
+    ENV_AUDIT_KAFKA_TLS_CA,
+    ENV_AUDIT_KAFKA_TLS_CLIENT_CERT,
+    ENV_AUDIT_KAFKA_TLS_CLIENT_KEY,
+    ENV_AUDIT_KAFKA_QUEUE_DIR,
+    ENV_AUDIT_KAFKA_QUEUE_LIMIT,
+];
+
+/// A list of all valid configuration keys for a Kafka audit target.
+pub const AUDIT_KAFKA_KEYS: &[&str] = &[
+    crate::ENABLE_KEY,
+    crate::KAFKA_BROKERS,
+    crate::KAFKA_TOPIC,
+    crate::KAFKA_ACKS,
+    crate::KAFKA_TLS_ENABLE,
+    crate::KAFKA_TLS_CA,
+    crate::KAFKA_TLS_CLIENT_CERT,
+    crate::KAFKA_TLS_CLIENT_KEY,
+    crate::KAFKA_QUEUE_DIR,
+    crate::KAFKA_QUEUE_LIMIT,
+    crate::COMMENT_KEY,
+];

crates/config/src/audit/mod.rs

@@ -16,14 +16,24 @@
 //! This module defines the configuration for audit systems, including
 //! webhook and MQTT audit-related settings.
 
+mod amqp;
+mod kafka;
 mod mqtt;
+mod mysql;
 mod nats;
+mod postgres;
 mod pulsar;
+mod redis;
 mod webhook;
 
+pub use amqp::*;
+pub use kafka::*;
 pub use mqtt::*;
+pub use mysql::*;
 pub use nats::*;
+pub use postgres::*;
 pub use pulsar::*;
+pub use redis::*;
 pub use webhook::*;
 
 use crate::DEFAULT_DELIMITER;
@@ -33,15 +43,24 @@ pub const AUDIT_PREFIX: &str = "audit";
 pub const AUDIT_ROUTE_PREFIX: &str = const_str::concat!(AUDIT_PREFIX, DEFAULT_DELIMITER);
 
 pub const AUDIT_WEBHOOK_SUB_SYS: &str = "audit_webhook";
+pub const AUDIT_AMQP_SUB_SYS: &str = "audit_amqp";
+pub const AUDIT_KAFKA_SUB_SYS: &str = "audit_kafka";
 pub const AUDIT_MQTT_SUB_SYS: &str = "audit_mqtt";
+pub const AUDIT_MYSQL_SUB_SYS: &str = "audit_mysql";
 pub const AUDIT_NATS_SUB_SYS: &str = "audit_nats";
+pub const AUDIT_POSTGRES_SUB_SYS: &str = "audit_postgres";
 pub const AUDIT_PULSAR_SUB_SYS: &str = "audit_pulsar";
-
+pub const AUDIT_REDIS_SUB_SYS: &str = "audit_redis";
+pub const AUDIT_REDIS_DEFAULT_CHANNEL: &str = "rustfs_audit_channel";
 pub const AUDIT_STORE_EXTENSION: &str = ".audit";
-#[allow(dead_code)]
 pub const AUDIT_SUB_SYSTEMS: &[&str] = &[
+    AUDIT_AMQP_SUB_SYS,
+    AUDIT_KAFKA_SUB_SYS,
     AUDIT_MQTT_SUB_SYS,
+    AUDIT_MYSQL_SUB_SYS,
     AUDIT_NATS_SUB_SYS,
+    AUDIT_POSTGRES_SUB_SYS,
     AUDIT_PULSAR_SUB_SYS,
+    AUDIT_REDIS_SUB_SYS,
     AUDIT_WEBHOOK_SUB_SYS,
 ];

crates/config/src/audit/mysql.rs

@@ -0,0 +1,53 @@
+// Copyright 2024 RustFS Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+/// A list of all valid configuration keys for an audit MySQL target.
+pub const AUDIT_MYSQL_KEYS: &[&str] = &[
+    crate::ENABLE_KEY,
+    crate::MYSQL_DSN_STRING,
+    crate::MYSQL_TABLE,
+    crate::MYSQL_FORMAT,
+    crate::MYSQL_TLS_CA,
+    crate::MYSQL_TLS_CLIENT_CERT,
+    crate::MYSQL_TLS_CLIENT_KEY,
+    crate::MYSQL_QUEUE_DIR,
+    crate::MYSQL_QUEUE_LIMIT,
+    crate::MYSQL_MAX_OPEN_CONNECTIONS,
+    crate::COMMENT_KEY,
+];
+
+// MySQL environment variables for audit target.
+pub const ENV_AUDIT_MYSQL_ENABLE: &str = "RUSTFS_AUDIT_MYSQL_ENABLE";
+pub const ENV_AUDIT_MYSQL_DSN_STRING: &str = "RUSTFS_AUDIT_MYSQL_DSN_STRING";
+pub const ENV_AUDIT_MYSQL_TABLE: &str = "RUSTFS_AUDIT_MYSQL_TABLE";
+pub const ENV_AUDIT_MYSQL_FORMAT: &str = "RUSTFS_AUDIT_MYSQL_FORMAT";
+pub const ENV_AUDIT_MYSQL_TLS_CA: &str = "RUSTFS_AUDIT_MYSQL_TLS_CA";
+pub const ENV_AUDIT_MYSQL_TLS_CLIENT_CERT: &str = "RUSTFS_AUDIT_MYSQL_TLS_CLIENT_CERT";
+pub const ENV_AUDIT_MYSQL_TLS_CLIENT_KEY: &str = "RUSTFS_AUDIT_MYSQL_TLS_CLIENT_KEY";
+pub const ENV_AUDIT_MYSQL_QUEUE_DIR: &str = "RUSTFS_AUDIT_MYSQL_QUEUE_DIR";
+pub const ENV_AUDIT_MYSQL_QUEUE_LIMIT: &str = "RUSTFS_AUDIT_MYSQL_QUEUE_LIMIT";
+pub const ENV_AUDIT_MYSQL_MAX_OPEN_CONNECTIONS: &str = "RUSTFS_AUDIT_MYSQL_MAX_OPEN_CONNECTIONS";
+
+pub const ENV_AUDIT_MYSQL_KEYS: &[&str; 10] = &[
+    ENV_AUDIT_MYSQL_ENABLE,
+    ENV_AUDIT_MYSQL_DSN_STRING,
+    ENV_AUDIT_MYSQL_TABLE,
+    ENV_AUDIT_MYSQL_FORMAT,
+    ENV_AUDIT_MYSQL_TLS_CA,
+    ENV_AUDIT_MYSQL_TLS_CLIENT_CERT,
+    ENV_AUDIT_MYSQL_TLS_CLIENT_KEY,
+    ENV_AUDIT_MYSQL_QUEUE_DIR,
+    ENV_AUDIT_MYSQL_QUEUE_LIMIT,
+    ENV_AUDIT_MYSQL_MAX_OPEN_CONNECTIONS,
+];

crates/config/src/audit/postgres.rs

@@ -0,0 +1,51 @@
+// Copyright 2024 RustFS Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+pub const AUDIT_POSTGRES_KEYS: &[&str] = &[
+    crate::ENABLE_KEY,
+    crate::POSTGRES_DSN_STRING,
+    crate::POSTGRES_TABLE,
+    crate::POSTGRES_FORMAT,
+    crate::POSTGRES_TLS_REQUIRED,
+    crate::POSTGRES_TLS_CA,
+    crate::POSTGRES_TLS_CLIENT_CERT,
+    crate::POSTGRES_TLS_CLIENT_KEY,
+    crate::POSTGRES_QUEUE_DIR,
+    crate::POSTGRES_QUEUE_LIMIT,
+    crate::COMMENT_KEY,
+];
+
+pub const ENV_AUDIT_POSTGRES_ENABLE: &str = "RUSTFS_AUDIT_POSTGRES_ENABLE";
+pub const ENV_AUDIT_POSTGRES_DSN_STRING: &str = "RUSTFS_AUDIT_POSTGRES_DSN_STRING";
+pub const ENV_AUDIT_POSTGRES_TABLE: &str = "RUSTFS_AUDIT_POSTGRES_TABLE";
+pub const ENV_AUDIT_POSTGRES_FORMAT: &str = "RUSTFS_AUDIT_POSTGRES_FORMAT";
+pub const ENV_AUDIT_POSTGRES_TLS_REQUIRED: &str = "RUSTFS_AUDIT_POSTGRES_TLS_REQUIRED";
+pub const ENV_AUDIT_POSTGRES_TLS_CA: &str = "RUSTFS_AUDIT_POSTGRES_TLS_CA";
+pub const ENV_AUDIT_POSTGRES_TLS_CLIENT_CERT: &str = "RUSTFS_AUDIT_POSTGRES_TLS_CLIENT_CERT";
+pub const ENV_AUDIT_POSTGRES_TLS_CLIENT_KEY: &str = "RUSTFS_AUDIT_POSTGRES_TLS_CLIENT_KEY";
+pub const ENV_AUDIT_POSTGRES_QUEUE_DIR: &str = "RUSTFS_AUDIT_POSTGRES_QUEUE_DIR";
+pub const ENV_AUDIT_POSTGRES_QUEUE_LIMIT: &str = "RUSTFS_AUDIT_POSTGRES_QUEUE_LIMIT";
+
+pub const ENV_AUDIT_POSTGRES_KEYS: &[&str; 10] = &[
+    ENV_AUDIT_POSTGRES_ENABLE,
+    ENV_AUDIT_POSTGRES_DSN_STRING,
+    ENV_AUDIT_POSTGRES_TABLE,
+    ENV_AUDIT_POSTGRES_FORMAT,
+    ENV_AUDIT_POSTGRES_TLS_REQUIRED,
+    ENV_AUDIT_POSTGRES_TLS_CA,
+    ENV_AUDIT_POSTGRES_TLS_CLIENT_CERT,
+    ENV_AUDIT_POSTGRES_TLS_CLIENT_KEY,
+    ENV_AUDIT_POSTGRES_QUEUE_DIR,
+    ENV_AUDIT_POSTGRES_QUEUE_LIMIT,
+];

crates/config/src/audit/redis.rs

@@ -0,0 +1,81 @@
+// Copyright 2024 RustFS Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+pub const ENV_AUDIT_REDIS_ENABLE: &str = "RUSTFS_AUDIT_REDIS_ENABLE";
+pub const ENV_AUDIT_REDIS_URL: &str = "RUSTFS_AUDIT_REDIS_URL";
+pub const ENV_AUDIT_REDIS_CHANNEL: &str = "RUSTFS_AUDIT_REDIS_CHANNEL";
+pub const ENV_AUDIT_REDIS_USERNAME: &str = "RUSTFS_AUDIT_REDIS_USERNAME";
+pub const ENV_AUDIT_REDIS_PASSWORD: &str = "RUSTFS_AUDIT_REDIS_PASSWORD";
+pub const ENV_AUDIT_REDIS_KEEP_ALIVE_INTERVAL: &str = "RUSTFS_AUDIT_REDIS_KEEP_ALIVE_INTERVAL";
+pub const ENV_AUDIT_REDIS_QUEUE_DIR: &str = "RUSTFS_AUDIT_REDIS_QUEUE_DIR";
+pub const ENV_AUDIT_REDIS_QUEUE_LIMIT: &str = "RUSTFS_AUDIT_REDIS_QUEUE_LIMIT";
+pub const ENV_AUDIT_REDIS_MAX_RETRY_ATTEMPTS: &str = "RUSTFS_AUDIT_REDIS_MAX_RETRY_ATTEMPTS";
+pub const ENV_AUDIT_REDIS_RECONNECT_RETRY_ATTEMPTS: &str = "RUSTFS_AUDIT_REDIS_RECONNECT_RETRY_ATTEMPTS";
+pub const ENV_AUDIT_REDIS_MIN_RETRY_DELAY: &str = "RUSTFS_AUDIT_REDIS_MIN_RETRY_DELAY";
+pub const ENV_AUDIT_REDIS_MAX_RETRY_DELAY: &str = "RUSTFS_AUDIT_REDIS_MAX_RETRY_DELAY";
+pub const ENV_AUDIT_REDIS_CONNECTION_TIMEOUT: &str = "RUSTFS_AUDIT_REDIS_CONNECTION_TIMEOUT";
+pub const ENV_AUDIT_REDIS_RESPONSE_TIMEOUT: &str = "RUSTFS_AUDIT_REDIS_RESPONSE_TIMEOUT";
+pub const ENV_AUDIT_REDIS_PIPELINE_BUFFER_SIZE: &str = "RUSTFS_AUDIT_REDIS_PIPELINE_BUFFER_SIZE";
+pub const ENV_AUDIT_REDIS_TLS_POLICY: &str = "RUSTFS_AUDIT_REDIS_TLS_POLICY";
+pub const ENV_AUDIT_REDIS_TLS_CA: &str = "RUSTFS_AUDIT_REDIS_TLS_CA";
+pub const ENV_AUDIT_REDIS_TLS_CLIENT_CERT: &str = "RUSTFS_AUDIT_REDIS_TLS_CLIENT_CERT";
+pub const ENV_AUDIT_REDIS_TLS_CLIENT_KEY: &str = "RUSTFS_AUDIT_REDIS_TLS_CLIENT_KEY";
+pub const ENV_AUDIT_REDIS_TLS_ALLOW_INSECURE: &str = "RUSTFS_AUDIT_REDIS_TLS_ALLOW_INSECURE";
+
+pub const ENV_AUDIT_REDIS_KEYS: &[&str; 20] = &[
+    ENV_AUDIT_REDIS_ENABLE,
+    ENV_AUDIT_REDIS_URL,
+    ENV_AUDIT_REDIS_CHANNEL,
+    ENV_AUDIT_REDIS_USERNAME,
+    ENV_AUDIT_REDIS_PASSWORD,
+    ENV_AUDIT_REDIS_KEEP_ALIVE_INTERVAL,
+    ENV_AUDIT_REDIS_QUEUE_DIR,
+    ENV_AUDIT_REDIS_QUEUE_LIMIT,
+    ENV_AUDIT_REDIS_MAX_RETRY_ATTEMPTS,
+    ENV_AUDIT_REDIS_RECONNECT_RETRY_ATTEMPTS,
+    ENV_AUDIT_REDIS_MIN_RETRY_DELAY,
+    ENV_AUDIT_REDIS_MAX_RETRY_DELAY,
+    ENV_AUDIT_REDIS_CONNECTION_TIMEOUT,
+    ENV_AUDIT_REDIS_RESPONSE_TIMEOUT,
+    ENV_AUDIT_REDIS_PIPELINE_BUFFER_SIZE,
+    ENV_AUDIT_REDIS_TLS_POLICY,
+    ENV_AUDIT_REDIS_TLS_CA,
+    ENV_AUDIT_REDIS_TLS_CLIENT_CERT,
+    ENV_AUDIT_REDIS_TLS_CLIENT_KEY,
+    ENV_AUDIT_REDIS_TLS_ALLOW_INSECURE,
+];
+
+pub const AUDIT_REDIS_KEYS: &[&str] = &[
+    crate::ENABLE_KEY,
+    crate::REDIS_URL,
+    crate::REDIS_CHANNEL,
+    crate::REDIS_USERNAME,
+    crate::REDIS_PASSWORD,
+    crate::REDIS_KEEP_ALIVE_INTERVAL,
+    crate::REDIS_QUEUE_DIR,
+    crate::REDIS_QUEUE_LIMIT,
+    crate::REDIS_MAX_RETRY_ATTEMPTS,
+    crate::REDIS_RECONNECT_RETRY_ATTEMPTS,
+    crate::REDIS_MIN_RETRY_DELAY,
+    crate::REDIS_MAX_RETRY_DELAY,
+    crate::REDIS_CONNECTION_TIMEOUT,
+    crate::REDIS_RESPONSE_TIMEOUT,
+    crate::REDIS_PIPELINE_BUFFER_SIZE,
+    crate::REDIS_TLS_POLICY,
+    crate::REDIS_TLS_CA,
+    crate::REDIS_TLS_CLIENT_CERT,
+    crate::REDIS_TLS_CLIENT_KEY,
+    crate::REDIS_TLS_ALLOW_INSECURE,
+    crate::COMMENT_KEY,
+];

crates/config/src/constants/app.rs

@@ -131,23 +131,45 @@ pub const ENV_RUSTFS_ADDRESS: &str = "RUSTFS_ADDRESS";
 /// Environment variable for server volumes.
 pub const ENV_RUSTFS_VOLUMES: &str = "RUSTFS_VOLUMES";
 
+/// Environment variable to explicitly bypass local physical disk independence checks.
+pub const ENV_UNSAFE_BYPASS_DISK_CHECK: &str = "RUSTFS_UNSAFE_BYPASS_DISK_CHECK";
+
+/// Compatibility alias used by legacy MinIO CI pipelines.
+///
+/// RustFS keeps this alias for backward compatibility only. Prefer
+/// `ENV_UNSAFE_BYPASS_DISK_CHECK` for explicit bypass control.
+pub const ENV_MINIO_CI: &str = "MINIO_CI";
+
+/// Default flag value for bypassing local physical disk independence checks.
+pub const DEFAULT_UNSAFE_BYPASS_DISK_CHECK: bool = false;
+
 /// Environment variable for server access key.
 pub const ENV_RUSTFS_ACCESS_KEY: &str = "RUSTFS_ACCESS_KEY";
 
 /// Environment variable for server access key file.
 pub const ENV_RUSTFS_ACCESS_KEY_FILE: &str = "RUSTFS_ACCESS_KEY_FILE";
 
-/// Environment variable for server root user.
-pub const ENV_RUSTFS_ROOT_USER: &str = "RUSTFS_ROOT_USER";
-
 /// Environment variable for server secret key.
 pub const ENV_RUSTFS_SECRET_KEY: &str = "RUSTFS_SECRET_KEY";
 
 /// Environment variable for server secret key file.
 pub const ENV_RUSTFS_SECRET_KEY_FILE: &str = "RUSTFS_SECRET_KEY_FILE";
 
-/// Environment variable for server root password.
-pub const ENV_RUSTFS_ROOT_PASSWORD: &str = "RUSTFS_ROOT_PASSWORD";
+/// Environment variable to explicitly allow public default root credentials.
+///
+/// This is intended for local development only. Production startup paths should
+/// provide non-default `RUSTFS_ACCESS_KEY` and `RUSTFS_SECRET_KEY` values.
+pub const ENV_RUSTFS_ALLOW_INSECURE_DEFAULT_CREDENTIALS: &str = "RUSTFS_ALLOW_INSECURE_DEFAULT_CREDENTIALS";
+
+/// Environment variable controlling startup behavior when unsupported filesystem types are detected.
+///
+/// Accepted values:
+/// - "warn" (default): log a warning and continue startup
+/// - "fail": abort startup with an error
+pub const ENV_RUSTFS_UNSUPPORTED_FS_POLICY: &str = "RUSTFS_UNSUPPORTED_FS_POLICY";
+pub const RUSTFS_UNSUPPORTED_FS_POLICY_WARN: &str = "warn";
+pub const RUSTFS_UNSUPPORTED_FS_POLICY_FAIL: &str = "fail";
+pub const DEFAULT_RUSTFS_UNSUPPORTED_FS_POLICY: &str = RUSTFS_UNSUPPORTED_FS_POLICY_WARN;
 
 /// Environment variable for server OBS endpoint.
 pub const ENV_RUSTFS_OBS_ENDPOINT: &str = "RUSTFS_OBS_ENDPOINT";
@@ -223,6 +245,9 @@ pub const ENV_RUSTFS_REGION: &str = "RUSTFS_REGION";
 /// Environment variable for server license.
 pub const ENV_RUSTFS_LICENSE: &str = "RUSTFS_LICENSE";
 
+/// Environment variable for the RSA public key used to verify server licenses.
+pub const ENV_RUSTFS_LICENSE_PUBLIC_KEY: &str = "RUSTFS_LICENSE_PUBLIC_KEY";
+
 /// Default log filename for rustfs
 /// This is the default log filename for rustfs.
 /// It is used to store the logs of the application.
@@ -336,6 +361,7 @@ mod tests {
     fn test_environment_constants() {
         // Test environment related constants
         assert_eq!(ENVIRONMENT, "production");
+        assert_eq!(ENV_RUSTFS_LICENSE_PUBLIC_KEY, "RUSTFS_LICENSE_PUBLIC_KEY");
         assert!(
             ["development", "staging", "production", "test"].contains(&ENVIRONMENT),
             "Environment should be a standard environment name"

crates/config/src/constants/console.rs

@@ -17,16 +17,21 @@
 pub const ENV_CORS_ALLOWED_ORIGINS: &str = "RUSTFS_CORS_ALLOWED_ORIGINS";
 
 /// Default CORS allowed origins for the endpoint service
-/// Comes from the console service default
-/// See DEFAULT_CONSOLE_CORS_ALLOWED_ORIGINS
-pub const DEFAULT_CORS_ALLOWED_ORIGINS: &str = DEFAULT_CONSOLE_CORS_ALLOWED_ORIGINS;
+/// Empty means the S3 endpoint emits no generic CORS headers unless configured.
+pub const DEFAULT_CORS_ALLOWED_ORIGINS: &str = "";
 
 /// CORS allowed origins for the console service
 /// Comma-separated list of origins or "*" for all origins
 pub const ENV_CONSOLE_CORS_ALLOWED_ORIGINS: &str = "RUSTFS_CONSOLE_CORS_ALLOWED_ORIGINS";
 
-/// Default CORS allowed origins for the console service
-pub const DEFAULT_CONSOLE_CORS_ALLOWED_ORIGINS: &str = "*";
+/// Default CORS allowed origins for the console service.
+///
+/// Empty string means same-origin only — no `Access-Control-Allow-Origin`
+/// header is emitted, so browsers will not allow cross-origin reads of
+/// console responses by default. Operators that need cross-origin access set
+/// `RUSTFS_CONSOLE_CORS_ALLOWED_ORIGINS` to a comma-separated allow-list, or
+/// to `*` to keep the previous permissive behavior.
+pub const DEFAULT_CONSOLE_CORS_ALLOWED_ORIGINS: &str = "";
 
 /// Enable or disable the console service
 pub const ENV_CONSOLE_ENABLE: &str = "RUSTFS_CONSOLE_ENABLE";
@@ -89,3 +94,20 @@ pub const ENV_UPDATE_CHECK: &str = "RUSTFS_CHECK_UPDATE";
 
 /// Default value for update toggle
 pub const DEFAULT_UPDATE_CHECK: bool = true;
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn endpoint_cors_default_is_restrictive() {
+        assert_eq!(ENV_CORS_ALLOWED_ORIGINS, "RUSTFS_CORS_ALLOWED_ORIGINS");
+        assert_eq!(DEFAULT_CORS_ALLOWED_ORIGINS, "");
+    }
+
+    #[test]
+    fn console_cors_default_is_same_origin_only() {
+        assert_eq!(ENV_CONSOLE_CORS_ALLOWED_ORIGINS, "RUSTFS_CONSOLE_CORS_ALLOWED_ORIGINS");
+        assert_eq!(DEFAULT_CONSOLE_CORS_ALLOWED_ORIGINS, "");
+    }
+}

crates/config/src/constants/drive.rs

@@ -39,6 +39,24 @@ pub const DEFAULT_DRIVE_WALKDIR_TIMEOUT_SECS: u64 = 5;
 pub const ENV_DRIVE_WALKDIR_STALL_TIMEOUT_SECS: &str = "RUSTFS_DRIVE_WALKDIR_STALL_TIMEOUT_SECS";
 pub const DEFAULT_DRIVE_WALKDIR_STALL_TIMEOUT_SECS: u64 = 5;
 
+/// Interval in seconds between active health probes for local and remote drives.
+pub const ENV_DRIVE_ACTIVE_CHECK_INTERVAL_SECS: &str = "RUSTFS_DRIVE_ACTIVE_CHECK_INTERVAL_SECS";
+pub const DEFAULT_DRIVE_ACTIVE_CHECK_INTERVAL_SECS: u64 = 15;
+
+/// Timeout in seconds for a single active health probe.
+pub const ENV_DRIVE_ACTIVE_CHECK_TIMEOUT_SECS: &str = "RUSTFS_DRIVE_ACTIVE_CHECK_TIMEOUT_SECS";
+pub const DEFAULT_DRIVE_ACTIVE_CHECK_TIMEOUT_SECS: u64 = 5;
+
+/// Timeout-to-health transition policy for drive operations.
+///
+/// Accepted values:
+/// - "mark_failure" (default): timeout marks failure and may transition runtime state.
+/// - "ignore_scanner": timeout does not mark failure for scanner-sensitive operations.
+pub const ENV_DRIVE_TIMEOUT_HEALTH_ACTION: &str = "RUSTFS_DRIVE_TIMEOUT_HEALTH_ACTION";
+pub const DRIVE_TIMEOUT_HEALTH_ACTION_MARK_FAILURE: &str = "mark_failure";
+pub const DRIVE_TIMEOUT_HEALTH_ACTION_IGNORE_SCANNER: &str = "ignore_scanner";
+pub const DEFAULT_DRIVE_TIMEOUT_HEALTH_ACTION: &str = DRIVE_TIMEOUT_HEALTH_ACTION_MARK_FAILURE;
+
 /// Number of consecutive failures before a suspect drive is classified as offline.
 pub const ENV_DRIVE_SUSPECT_FAILURE_THRESHOLD: &str = "RUSTFS_DRIVE_SUSPECT_FAILURE_THRESHOLD";
 pub const DEFAULT_DRIVE_SUSPECT_FAILURE_THRESHOLD: u64 = 2;
@@ -58,3 +76,19 @@ pub const DEFAULT_DRIVE_OFFLINE_GRACE_PERIOD_SECS: u64 = 30;
 /// Duration in seconds after which a recovered drive is classified as long offline.
 pub const ENV_DRIVE_LONG_OFFLINE_THRESHOLD_SECS: &str = "RUSTFS_DRIVE_LONG_OFFLINE_THRESHOLD_SECS";
 pub const DEFAULT_DRIVE_LONG_OFFLINE_THRESHOLD_SECS: u64 = 172_800;
+
+/// Drive timeout profile preset.
+///
+/// Accepted values:
+/// - "default": keep current timeout defaults.
+/// - "high_latency": use a higher default timeout preset for scanner-sensitive and metadata operations.
+///
+/// Explicit per-operation overrides (`RUSTFS_DRIVE_*_TIMEOUT_SECS`) still take precedence.
+pub const ENV_DRIVE_TIMEOUT_PROFILE: &str = "RUSTFS_DRIVE_TIMEOUT_PROFILE";
+pub const DRIVE_TIMEOUT_PROFILE_DEFAULT: &str = "default";
+pub const DRIVE_TIMEOUT_PROFILE_HIGH_LATENCY: &str = "high_latency";
+pub const DEFAULT_DRIVE_TIMEOUT_PROFILE: &str = DRIVE_TIMEOUT_PROFILE_DEFAULT;
+
+/// Timeout preset (seconds) used when `RUSTFS_DRIVE_TIMEOUT_PROFILE=high_latency`
+/// and no per-operation timeout override is provided.
+pub const DRIVE_TIMEOUT_PROFILE_HIGH_LATENCY_SECS: u64 = 60;

crates/config/src/constants/env.rs

@@ -26,6 +26,21 @@ pub const RUSTFS_WEBHOOK_SKIP_TLS_VERIFY_DEFAULT: bool = false;
 pub const ENABLE_KEY: &str = "enable";
 pub const COMMENT_KEY: &str = "comment";
 
+/// Global switch for enabling the audit module.
+pub const ENV_AUDIT_ENABLE: &str = "RUSTFS_AUDIT_ENABLE";
+/// Global switch for enabling the notify module.
+pub const ENV_NOTIFY_ENABLE: &str = "RUSTFS_NOTIFY_ENABLE";
+/// Default global audit switch (disabled by default).
+pub const DEFAULT_AUDIT_ENABLE: bool = false;
+/// Default global notify switch (disabled by default).
+pub const DEFAULT_NOTIFY_ENABLE: bool = false;
+/// Canonical ILM process boundary env key (seconds).
+pub const ENV_ILM_PROCESS_TIME: &str = "RUSTFS_ILM_PROCESS_TIME";
+/// Deprecated ILM process boundary env key kept for compatibility.
+pub const ENV_ILM_PROCESS_TIME_DEPRECATED: &str = "_RUSTFS_ILM_PROCESS_TIME";
+/// Default ILM process boundary in seconds (24h).
+pub const DEFAULT_ILM_PROCESS_TIME_SECS: i32 = 86400;
+
 /// Medium-drawn lines separator
 /// This is used to separate words in environment variable names.
 pub const ENV_WORD_DELIMITER_DASH: &str = "-";
@@ -290,4 +305,10 @@ mod tests {
             assert!(state.is_disabled());
         }
     }
+
+    #[test]
+    fn test_global_audit_notify_switch_constants() {
+        assert_eq!(ENV_AUDIT_ENABLE, "RUSTFS_AUDIT_ENABLE");
+        assert_eq!(ENV_NOTIFY_ENABLE, "RUSTFS_NOTIFY_ENABLE");
+    }
 }

crates/config/src/constants/health.rs

@@ -0,0 +1,28 @@
+// Copyright 2024 RustFS Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+/// Enable or disable public `/health` and `/health/ready` endpoints.
+/// When disabled, the routes are not registered and return 404.
+pub const ENV_HEALTH_ENDPOINT_ENABLE: &str = "RUSTFS_HEALTH_ENDPOINT_ENABLE";
+pub const DEFAULT_HEALTH_ENDPOINT_ENABLE: bool = true;
+
+/// Cache TTL for storage readiness runtime-state evaluation (milliseconds).
+/// This reduces storage-layer pressure when probes are called at high frequency.
+pub const ENV_HEALTH_READINESS_CACHE_TTL_MS: &str = "RUSTFS_HEALTH_READINESS_CACHE_TTL_MS";
+pub const DEFAULT_HEALTH_READINESS_CACHE_TTL_MS: u64 = 1000;
+
+/// Enable minimal health payload mode for GET `/health*` responses.
+/// When enabled, only `status` and `ready` fields are returned.
+pub const ENV_HEALTH_MINIMAL_RESPONSE_ENABLE: &str = "RUSTFS_HEALTH_MINIMAL_RESPONSE_ENABLE";
+pub const DEFAULT_HEALTH_MINIMAL_RESPONSE_ENABLE: bool = false;

crates/config/src/constants/internode.rs

@@ -0,0 +1,76 @@
+// Copyright 2024 RustFS Team
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+/// Timeout for establishing a new internode gRPC connection.
+pub const ENV_INTERNODE_CONNECT_TIMEOUT_SECS: &str = "RUSTFS_INTERNODE_CONNECT_TIMEOUT_SECS";
+pub const DEFAULT_INTERNODE_CONNECT_TIMEOUT_SECS: u64 = 3;
+
+/// TCP keepalive interval for internode gRPC channels.
+pub const ENV_INTERNODE_TCP_KEEPALIVE_SECS: &str = "RUSTFS_INTERNODE_TCP_KEEPALIVE_SECS";
+pub const DEFAULT_INTERNODE_TCP_KEEPALIVE_SECS: u64 = 10;
+
+/// HTTP/2 keepalive interval for internode gRPC channels.
+pub const ENV_INTERNODE_HTTP2_KEEPALIVE_INTERVAL_SECS: &str = "RUSTFS_INTERNODE_HTTP2_KEEPALIVE_INTERVAL_SECS";
+pub const DEFAULT_INTERNODE_HTTP2_KEEPALIVE_INTERVAL_SECS: u64 = 5;
+
+/// HTTP/2 keepalive timeout for internode gRPC channels.
+pub const ENV_INTERNODE_HTTP2_KEEPALIVE_TIMEOUT_SECS: &str = "RUSTFS_INTERNODE_HTTP2_KEEPALIVE_TIMEOUT_SECS";
+pub const DEFAULT_INTERNODE_HTTP2_KEEPALIVE_TIMEOUT_SECS: u64 = 3;
+
+/// Overall timeout for a single internode gRPC request.
+pub const ENV_INTERNODE_RPC_TIMEOUT_SECS: &str = "RUSTFS_INTERNODE_RPC_TIMEOUT_SECS";
+pub const DEFAULT_INTERNODE_RPC_TIMEOUT_SECS: u64 = 10;
+
+/// Environment variable for selecting the internode data-plane transport backend.
+pub const ENV_RUSTFS_INTERNODE_DATA_TRANSPORT: &str = "RUSTFS_INTERNODE_DATA_TRANSPORT";
+pub const DEFAULT_INTERNODE_DATA_TRANSPORT: &str = "tcp-http";
+
+/// Legacy alias for "tcp-http". Both values select the TCP/HTTP transport backend.
+pub const INTERNODE_DATA_TRANSPORT_TCP: &str = "tcp";
+
+/// Known internode transport backend names accepted by the config parser.
+pub const KNOWN_INTERNODE_DATA_TRANSPORT_BACKENDS: &[&str] = &[DEFAULT_INTERNODE_DATA_TRANSPORT, INTERNODE_DATA_TRANSPORT_TCP];
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn internode_timeout_defaults_stay_in_expected_bounds() {
+        assert_eq!(DEFAULT_INTERNODE_CONNECT_TIMEOUT_SECS, 3);
+        assert_eq!(DEFAULT_INTERNODE_TCP_KEEPALIVE_SECS, 10);
+        assert_eq!(DEFAULT_INTERNODE_HTTP2_KEEPALIVE_INTERVAL_SECS, 5);
+        assert_eq!(DEFAULT_INTERNODE_HTTP2_KEEPALIVE_TIMEOUT_SECS, 3);
+        assert_eq!(DEFAULT_INTERNODE_RPC_TIMEOUT_SECS, 10);
+    }
+
+    #[test]
+    fn internode_timeout_env_names_are_stable() {
+        assert_eq!(ENV_INTERNODE_CONNECT_TIMEOUT_SECS, "RUSTFS_INTERNODE_CONNECT_TIMEOUT_SECS");
+        assert_eq!(ENV_INTERNODE_TCP_KEEPALIVE_SECS, "RUSTFS_INTERNODE_TCP_KEEPALIVE_SECS");
+        assert_eq!(
+            ENV_INTERNODE_HTTP2_KEEPALIVE_INTERVAL_SECS,
+            "RUSTFS_INTERNODE_HTTP2_KEEPALIVE_INTERVAL_SECS"
+        );
+        assert_eq!(
+            ENV_INTERNODE_HTTP2_KEEPALIVE_TIMEOUT_SECS,
+            "RUSTFS_INTERNODE_HTTP2_KEEPALIVE_TIMEOUT_SECS"
+        );
+        assert_eq!(ENV_INTERNODE_RPC_TIMEOUT_SECS, "RUSTFS_INTERNODE_RPC_TIMEOUT_SECS");
+        assert_eq!(ENV_RUSTFS_INTERNODE_DATA_TRANSPORT, "RUSTFS_INTERNODE_DATA_TRANSPORT");
+        assert_eq!(DEFAULT_INTERNODE_DATA_TRANSPORT, "tcp-http");
+        assert_eq!(INTERNODE_DATA_TRANSPORT_TCP, "tcp");
+        assert_eq!(KNOWN_INTERNODE_DATA_TRANSPORT_BACKENDS, &["tcp-http", "tcp"]);
+    }
+}

crates/config/src/constants/mod.rs

@@ -20,6 +20,8 @@ pub(crate) mod console;
 pub(crate) mod drive;
 pub(crate) mod env;
 pub(crate) mod heal;
+pub(crate) mod health;
+pub(crate) mod internode;
 pub(crate) mod object;
 pub(crate) mod oidc;
 pub(crate) mod profiler;

crates/config/src/constants/protocols.rs

@@ -57,3 +57,114 @@ pub const ENV_WEBDAV_CERTS_DIR: &str = "RUSTFS_WEBDAV_CERTS_DIR";
 pub const ENV_WEBDAV_CA_FILE: &str = "RUSTFS_WEBDAV_CA_FILE";
 pub const ENV_WEBDAV_MAX_BODY_SIZE: &str = "RUSTFS_WEBDAV_MAX_BODY_SIZE";
 pub const ENV_WEBDAV_REQUEST_TIMEOUT: &str = "RUSTFS_WEBDAV_REQUEST_TIMEOUT";
+
+/// Default SFTP server bind address.
+pub const DEFAULT_SFTP_ADDRESS: &str = "0.0.0.0:2222";
+
+/// Default for SFTP host-key directory. None means no default. Operators
+/// must set RUSTFS_SFTP_HOST_KEY_DIR explicitly when SFTP is enabled.
+pub const DEFAULT_SFTP_HOST_KEY_DIR: Option<&str> = None;
+
+/// SFTP environment variable names.
+pub const ENV_SFTP_ENABLE: &str = "RUSTFS_SFTP_ENABLE";
+pub const ENV_SFTP_ADDRESS: &str = "RUSTFS_SFTP_ADDRESS";
+pub const ENV_SFTP_HOST_KEY_DIR: &str = "RUSTFS_SFTP_HOST_KEY_DIR";
+pub const ENV_SFTP_HOST_KEY_RELOAD_ENABLE: &str = "RUSTFS_SFTP_HOST_KEY_RELOAD_ENABLE";
+pub const ENV_SFTP_HOST_KEY_RELOAD_INTERVAL: &str = "RUSTFS_SFTP_HOST_KEY_RELOAD_INTERVAL";
+pub const ENV_SFTP_IDLE_TIMEOUT: &str = "RUSTFS_SFTP_IDLE_TIMEOUT";
+/// S3 multipart part size in bytes. Default DEFAULT_SFTP_PART_SIZE (16 MiB).
+/// Valid range 5 MiB to 5 GiB (S3 protocol bounds), enforced at startup.
+///
+/// The per-upload size ceiling is part_size * 10_000 (the S3 parts cap),
+/// so the default caps single uploads at 160 GiB. Deployments expecting
+/// larger single files must raise this: 64 MiB -> 640 GiB, 128 MiB ->
+/// 1.25 TiB, 512 MiB -> 5 TiB (S3 object max). Rename is not affected;
+/// multipart_copy scales the per-part size dynamically and handles up
+/// to the 5 TiB S3 object limit regardless of this setting.
+pub const ENV_SFTP_PART_SIZE: &str = "RUSTFS_SFTP_PART_SIZE";
+pub const ENV_SFTP_READ_ONLY: &str = "RUSTFS_SFTP_READ_ONLY";
+pub const ENV_SFTP_BANNER: &str = "RUSTFS_SFTP_BANNER";
+/// Optional environment variable. If RUSTFS_SFTP_HANDLES_PER_SESSION
+/// is not set in the process environment, the server uses the default
+/// of 64 handles per session and emits no warning. If set, the value
+/// must be in the inclusive range 8 to 1024. Out-of-range values fall
+/// back to the default of 64 with a warn-level log naming the
+/// requested value and the bounds.
+///
+/// Caps the maximum number of simultaneously-open SFTP handles per
+/// session. A handle is the server-side identifier returned by
+/// SSH_FXP_OPEN and SSH_FXP_OPENDIR. One client typically uses one
+/// handle per file in flight plus one per directory listing.
+/// Operators running clients with deep pipelining may raise this.
+pub const ENV_SFTP_HANDLES_PER_SESSION: &str = "RUSTFS_SFTP_HANDLES_PER_SESSION";
+
+/// Optional environment variable. If RUSTFS_SFTP_BACKEND_OP_TIMEOUT_SECS
+/// is not set in the process environment, the server uses the default
+/// of 60 seconds and emits no warning. If set, the value must be in
+/// the inclusive range 5 to 600 seconds. Out-of-range values fall
+/// back to the default with a warn-level log naming the requested
+/// value and the bounds.
+///
+/// Bounds every storage backend call issued by the SFTP driver. A
+/// backend that does not respond within this many seconds returns
+/// Failure to the client and emits a warn log naming the backend
+/// method. This catches a backend that accepted the request and never
+/// returned a body, which the SSH keepalive cannot detect because the
+/// transport itself remains live.
+pub const ENV_SFTP_BACKEND_OP_TIMEOUT_SECS: &str = "RUSTFS_SFTP_BACKEND_OP_TIMEOUT_SECS";
+
+/// Optional environment variable. If RUSTFS_SFTP_READ_CACHE_WINDOW_BYTES
+/// is not set in the process environment, the server uses a 4 MiB
+/// default and emits no warning. If set, the value must be in the
+/// inclusive range MAX_READ_LEN (256 KiB) to 64 MiB. Out-of-range
+/// values fall back to the default with a warn-level log naming the
+/// requested value and the bounds.
+///
+/// Per-handle byte window the SFTP read path fetches in one backend
+/// call on a cache miss. Subsequent FXP_READs within that window are
+/// served from the buffer without a backend round trip. For
+/// sequential downloads the backend round-trip count drops by
+/// window_bytes / MAX_READ_LEN. Random-access workloads should set
+/// the window equal to MAX_READ_LEN to opt out of read-ahead.
+pub const ENV_SFTP_READ_CACHE_WINDOW_BYTES: &str = "RUSTFS_SFTP_READ_CACHE_WINDOW_BYTES";
+
+/// Optional environment variable. If
+/// RUSTFS_SFTP_READ_CACHE_TOTAL_MEM_BYTES is not set in the process
+/// environment, the server uses a 256 MiB default and emits no
+/// warning. If set, the value must be at least 16 MiB. Below-min
+/// values fall back to the default with a warn-level log naming the
+/// requested value and the bound.
+///
+/// Process-wide ceiling on cumulative read cache memory across every
+/// live SFTP handle. Once the accumulator plus a new window would
+/// exceed this value, the populate call on the per-handle cache is
+/// skipped. The read still completes from the freshly-fetched bytes
+/// without storing them in the cache, at the cost of one backend
+/// call per FXP_READ. High-concurrency deployments expecting many
+/// parallel downloads should raise this in step with the per-session
+/// handle cap.
+pub const ENV_SFTP_READ_CACHE_TOTAL_MEM_BYTES: &str = "RUSTFS_SFTP_READ_CACHE_TOTAL_MEM_BYTES";
+
+/// Default idle session timeout in seconds.
+pub const DEFAULT_SFTP_IDLE_TIMEOUT: u64 = 600;
+
+/// Default SFTP host key hot reload enabled state.
+pub const DEFAULT_SFTP_HOST_KEY_RELOAD_ENABLE: bool = false;
+
+/// Default SFTP host key hot reload interval in seconds.
+pub const DEFAULT_SFTP_HOST_KEY_RELOAD_INTERVAL: u64 = 30;
+
+/// Default S3 multipart upload part size in bytes (16 MiB).
+///
+/// The per-upload size ceiling is part_size * 10_000 (the S3 parts cap),
+/// so the default gives a 160 GiB single-upload limit. Deployments that
+/// expect single files larger than this must raise part_size:
+/// 64 MiB -> 640 GiB, 128 MiB -> 1.25 TiB, 512 MiB -> 5 TiB (S3 max).
+/// The minimum is 5 MiB and the maximum is 5 GiB (S3 protocol bounds).
+pub const DEFAULT_SFTP_PART_SIZE: u64 = 16_777_216;
+
+/// Default read-only mode (disabled).
+pub const DEFAULT_SFTP_READ_ONLY: bool = false;
+
+/// Default SSH identification string (no version disclosure).
+pub const DEFAULT_SFTP_BANNER: &str = "SSH-2.0-RustFS";