Commit Graph

4981 Commits

Author SHA1 Message Date
pulse-triage[bot] 7331f3579d Merge reviewed scoped agent identity lookups
Preserve exact Core candidate history and complete published monitoring inventory while removing unrelated temporary metadata copies.

Change-source: pulse-maintainer
2026-10-05 02:32:33 +01:00
pulse-triage[bot] 36fc9fbf89 Make agent copy-ownership expectations independent
Freeze the source expectation as serialized bytes before mutating returned records. A broken clone cannot change both the tested state and its reference, so module, update and applied-config alias regressions must be detected.

Change-source: pulse-maintainer
2026-10-05 01:45:52 +01:00
pulse-triage[bot] 39906a220a Compare agent lookup evidence independently of registry build time
The unchanged-parent allocation controls fail as intended. Its semantic oracle also compared separate wall-clock registry construction timestamps. Retain all source timestamps and canonical payloads while excluding only Resource.UpdatedAt from the full-reference comparison.

Change-source: pulse-maintainer
2026-10-05 01:22:51 +01:00
pulse-triage[bot] 187e5daaa8 Keep agent identity lookups scoped to relevant inventory
Avoid cloning and ingesting twenty known non-host Kubernetes metadata kinds in temporary per-report matching registries, while preserving all publication and canonical matching. Clone only the selected native record for exact host reads and detach mutable host update, module and config state. Pin reference equivalence, live report publication, ownership and allocation scaling with connected controls.

Change-source: pulse-maintainer
2026-10-05 01:17:08 +01:00
pulse-triage[bot] bedb245c72 Reproduce fleet-sized work in agent identity lookups
Compare complete canonical host views with a full-snapshot reference and pin current replacement. Measure unrelated Kubernetes metadata and peer-host allocation growth before repairing report lookups.

Change-source: pulse-maintainer
2026-10-05 01:08:54 +01:00
pulse-triage[bot] f1c7e413ba Integrate published demo workload names into reviewed guest safety batch
Preserve every accepted commit and compose the exact current published frontier before publication.

Change-source: pulse-maintainer
2026-10-05 01:04:27 +01:00
pulse-triage[bot] 8ccb82a7f9 Merge reviewed platform-aware guest Discovery verification
Resolve the prior non-Linux skip in favor of the exact reviewed real-registration refusal fixture. Preserve candidate a2608c9 and runtime safety contracts.

Change-source: pulse-maintainer
2026-10-05 00:28:07 +01:00
pulse-triage[bot] 1f0d12af85 Match safety fixtures to private responses and native platforms
PR2472 build run 37239183545 extracts no wrapped TrueNAS recipes; exercise all three current blocks and retain synthetic HTTP response bodies only in private files, including 201 save and error controls. Native run 37239183481 expects a Linux QGA capability on macOS and Windows; assert non-Linux agents do not advertise it and scope the connected PVE fixture to its actual Linux runtime. Production safety and documentation remain unchanged.

Contract-Neutral: Test-only reconciliation of already-reviewed private-response recipes and existing Linux-only guest-execution capability; no product contract change.
Change-source: pulse-maintainer
2026-10-05 00:08:47 +01:00
rcourtman b8554bdd78 Stop repeating Kubernetes workload names in the demo
The demo scenario renamed every generated deployment to one of four
curated names and every pod to one of ten, cycling by index. With the
default 14 deployments and 40 pods per cluster the Kubernetes overview
listed checkout-api four times in one namespace, and the curated
pending and crash-looping pods appeared more than once. Kubernetes
itself refuses duplicate names in a namespace, so the public demo
looked broken.

Ten more healthy services fill out the default deployment count, a
numeric suffix keeps any larger count unique, and pods past the curated
story become healthy ReplicaSet-style replicas of the cluster's own
deployments on ready nodes, so each curated failure appears once.

Contract-Neutral: Mock fixture naming only: demo Kubernetes deployments and pods get unique names; no runtime contract delta.
2026-10-05 00:00:14 +01:00
pulse-triage[bot] f1e221a0fb Keep failed guest metadata refreshes from renewing evidence
Retain metadata values and the last useful observation time through failed, empty and unusable refreshes. Apply the existing retry backoff separately instead of making retained identity new guest-agent evidence or a fresh complete-cache hit. Preserve useful partial responses, normal cache hits and OS failure policy, with HTTP polling and recovery controls.

Contract-Neutral: Behavioural correction to existing guest metadata observation timestamps and retry scheduling; no schema, API, route, field, guest command, admission, duration, retention or numeric source-selection contract delta.
Change-source: pulse-maintainer
2026-10-04 23:34:19 +01:00
pulse-triage[bot] a2608c9fad Match guest Discovery integration proof to native platform capabilities
Native Windows and macOS correctly advertise no PVE guest guard, so their real-server fixture must assert refusal before any config or QGA handoff. Keep the existing Linux admission, outcome and resumption assertions and exercise foreign platform labels without changing runtime capability policy.

Contract-Neutral: Test-only platform evidence repair; guest execution and Discovery runtime contracts, capability advertisement and safety gates are unchanged.
Change-source: pulse-maintainer
2026-10-04 22:49:53 +01:00
pulse-triage[bot] a15920eaeb Keep retained guest observations out of new History points
Use the existing producer-owned memory observation and filesystem validity in the native Proxmox History writer. Preserve old points and legacy unannotated policy while keeping independent current memory, CPU, valid I/O and measured-zero recovery. Controls reproduce the unchanged-parent failures through real HTTP polling, in-memory and persisted series, and short/long chart reads.

Contract-Neutral: Behavioural repair consuming existing guest observation and filesystem validity in the existing History path; no schema, API, route, numeric source selection, guest-agent admission, alert threshold, timestamp precision or retention contract delta.
Change-source: pulse-maintainer
2026-10-04 22:43:53 +01:00
pulse-triage[bot] 0fd7768e9d Preserve explicit filesystem disablement during deferred reads
Separate operator policy from retained numeric evidence. Apply a filesystem disable override without evaluating its old usage or treating absent deferred inventory as authoritative removal.

Change-source: pulse-maintainer
2026-10-04 21:41:32 +01:00
pulse-triage[bot] 5ce63f1400 Keep retained guest readings out of alert decisions
Use selected memory observation state and filesystem read validity at both threshold and filter boundaries. Retained or failed inventory cannot open or resolve incidents, refresh their evidence or remove per-filesystem alerts. Independent current metrics, legacy readings and deliberate disablement retain their existing policy.

Change-source: pulse-maintainer
2026-10-04 21:39:16 +01:00
pulse-triage[bot] b0f8b96c8a Reproduce retained guest readings changing alert state
Exercise production memory/filesystem alert admission, delayed breaches, filters, canonical recovery and the connected HTTP backup-lock polling boundary. Retained evidence is not a new observation or an authoritative filesystem removal.

Change-source: pulse-maintainer
2026-10-04 21:37:11 +01:00
pulse-triage[bot] c0f5e98d5d Exercise every CORE recovery regression poll cycle
A zero test deadline reinstates the real connection interval, so it did not make subsequent cycles due. Use a nonzero expired fixture deadline and assert each cycle reaches the disk endpoint; keep production scheduling and failure/recovery assertions unchanged.

Change-source: pulse-maintainer
2026-10-04 20:55:13 +01:00
pulse-triage[bot] b0c5eb1a82 Avoid full resource copies in metadata read helpers
Borrow the already detached resource for read-only sensitivity and platform facet scans. Keep the classification algorithm, scope rules and timed benchmark workloads, while verifying selected memory provenance and clone ownership.

Contract-Neutral: Internal read-only metadata borrowing preserves canonical scope, policy, clone, memory provenance and wire semantics.
Change-source: pulse-maintainer
2026-10-04 20:46:52 +01:00
pulse-triage[bot] 1ab97537d7 Keep CORE inventory when disk rotation metadata is unknown
Treat optional REST rotation rates as the RPC reader does, without weakening required inventory decoding or fabricating disk health. Cover complete CORE snapshots, mixed alert arguments, topology and saved-config poll failure/recovery/disable.

Refs: #2466
Change-source: pulse-maintainer
2026-10-04 20:42:02 +01:00
pulse-triage[bot] d7ef00f2f2 Keep VM command and Discovery evidence truthful
Project verified guest status and decoded output instead of qm transport success. Reject ambiguous terminal outcomes without replay; retain known failed or truncated results as failures without inventing uncertainty. Admit only successful stdout to Discovery and preserve saved results when attempted probes yield no evidence. Keep backup precautions, source bounds and native acceptance dependencies unchanged.

Change-source: pulse-maintainer
2026-10-04 19:46:32 +01:00
pulse-triage[bot] 4188c2a9ba Reproduce false VM command and Discovery success
Add source controls for actual guest exit/output projection, ambiguous terminal receipts, failed probe evidence exclusion, and saved manual/background Discovery through explicit recovery. The fixtures never invoke PVE or QGA.

Contract-Neutral: Test-only regression controls exercise existing command and Discovery boundaries without changing production code, payloads, capabilities or budgets.
Change-source: pulse-maintainer
2026-10-04 19:35:32 +01:00
pulse-triage[bot] 8b6513581c Align Discovery controls with explicit VM cancellation truth
Keep the generic catalogue-error fixture distinct from unknown timed-out guest handoff, and require the existing canceled-VM control to expose a safety deferral without dispatch. The new completion/cancellation parent failures and full-source adverse result remain retained; no production behavior changes in this commit.

Contract-Neutral: Test-only correction records the existing guest cancellation and ordinary-command error contracts without changing production source, APIs, protocol or time budgets.
Change-source: pulse-maintainer
2026-10-04 18:47:43 +01:00
pulse-triage[bot] f9ed9311f2 Keep interrupted VM Discovery as an explicit safety deferral
Cancellation before dispatch or server-side expiry before a final agent result must not turn optional VM probes into completed fresh evidence or saved metadata fallback. Preserve the original error and stop subsequent probes while the agent retains its independent uncertainty fence.

Contract-Neutral: Completes the existing Discovery unavailable-evidence and cancellation requirements without changing API identity, statuses, protocol, command catalogue, approvals or non-VM execution.
Change-source: pulse-maintainer
2026-10-04 18:35:15 +01:00
pulse-triage[bot] 5f43619505 Make URL security controls deterministic in offline source checks
The connected suite exposed three public-host success cases that depended on external DNS. Resolve fixture-only public, private, mixed and missing names through in-process DNS responses while exercising the real parser/resolver and all existing SSRF checks; production resolution and its failures are unchanged.

Contract-Neutral: Test-only deterministic DNS fixtures preserve production URL resolution, SSRF address policy, fetch transport and public interfaces.
Change-source: pulse-maintainer
2026-10-04 18:27:05 +01:00
pulse-triage[bot] 419d80e2e4 Require terminal guest status before releasing VM execution admission
A zero qm CLI exit may return only a handed-off PID or an unverified guest result. Preserve actual CLI output/exit semantics but pause uncertain terminal dictionaries without another QGA command; accept complete canonical termination evidence and keep native thaw acceptance separate.

Change-source: pulse-maintainer
2026-10-04 18:20:25 +01:00
pulse-triage[bot] 2a734a5c53 Keep canceled VM lock evidence an explicit execution deferral
A cancellation before or during postflight config verification must not become an ordinary optional Discovery failure with usable output, especially on the final probe. Preserve fixed safety classification and no-handoff truth on canceled local admission.

Contract-Neutral: Completes the existing cancellation and unavailable-evidence requirements without changing protocol fields, guard version, lock/identity admission, time budgets or command catalogues.
Change-source: pulse-maintainer
2026-10-04 18:15:30 +01:00
pulse-triage[bot] 713e8d9e4b Keep guest-safety fixtures on canonical Discovery identities
Retain nested Docker VM scanner deferral controls without treating that execution-only type as a public Discovery/store identity. Exercise saved/manual/background preservation on canonical VM requests; runtime and API scope are unchanged.

Change-source: pulse-maintainer
2026-10-04 18:11:55 +01:00
pulse-triage[bot] d20b15b1e6 Complete Discovery compilation and accept canonical PVE config keys
The first exact proof established all ten parent no-admission failures and registered-session controls, then stopped on a missing strings import. Complete that import and preserve ordinary hyphenated PVE dictionary keys while keeping duplicate/case-conflicting lock rejection.

Contract-Neutral: Preserves the guest safety protocol and its lock, identity, budget and dispatch requirements; completes compilation and accepts canonical hyphenated config keys.
Change-source: pulse-maintainer
2026-10-04 18:05:57 +01:00
pulse-triage[bot] be319e8204 Guard VM Discovery and command execution at the owning host agent
Require local current-node lock evidence and registered guard support before QGA execution. Serialise VM probes, stop on deferral, retain actual in-flight outcomes and saved Discovery evidence, and pause uncertain completion without replay. Config checks do not reserve QGA or prove native thaw; the existing backup monitoring-outage precaution remains.

Change-source: pulse-maintainer
2026-10-04 17:59:09 +01:00
pulse-triage[bot] def087ed90 Reproduce unverified VM command admission at the host agent
Approvals and trusted Discovery commands currently cross QGA handoff without evidence of a local unlocked guest. Assert explicit no-handoff deferral without contacting a native target.

Change-source: pulse-maintainer
2026-10-04 17:41:40 +01:00
pulse-triage[bot] 80d8ad0f59 Keep unknown linked-agent observation times explicitly last-known
An online source status without a valid original receipt time cannot assert current memory freshness. Preserve its numeric availability and existing selection while omitting unknown/future time; cover both cases at the linked view. Remove the stray adapter-test comment copied during test relocation.

Contract-Neutral: Completes the optional memory-observation semantics and wire shape documented with 43ccc6c39a; no new contract or selection delta.
Change-source: pulse-maintainer
2026-10-04 17:03:47 +01:00
pulse-triage[bot] 43ccc6c39a Keep guest-memory provenance bound to its original observation
Expose current, last-known and unavailable memory through raw guest and selected unified metrics without renewing QGA evidence during repeated backup deferrals. Preserve independent PVE/linked-agent source times, numeric selection, History and alert policy. Annotate selected metrics as well as raw facets so cross-source fallback cannot inherit another source freshness. Add connected observation, source ownership, clone and wire controls and update the shared subsystem contracts.

Change-source: pulse-maintainer
2026-10-04 17:00:24 +01:00
pulse-triage[bot] c50b8340fd Initialise the guest-memory control metadata limiter
The first control panicked at an uninitialised synthetic Monitor map before reaching the provenance assertions. Supply the same map as production and retain that failed attempt without claiming a product regression.

Change-source: pulse-maintainer
2026-10-04 16:50:29 +01:00
pulse-triage[bot] abee023634 Reproduce lost guest-memory provenance through served projections
Exercise normal cache reads, repeated backup deferrals, expired evidence, source-independent PVE memory and resumption through the shared guest builder and real unified broadcast conversion. These are synthetic monitoring controls, not native thaw acceptance.

Change-source: pulse-maintainer
2026-10-04 16:45:34 +01:00
pulse-triage[bot] 7a19a06631 Keep backup inventory guard bound to the shared evaluator
The runtime preserves canonical backup inventory scope in recovery_rollups.go; update the existing ownership guard to follow that call and require both PVE/PBS callers rather than the former duplicated literal.

Change-source: pulse-maintainer
2026-10-04 16:01:54 +01:00
pulse-triage[bot] 5774b6bc4b Keep backup-age evaluation fast and failures visible
For #2465, select latest provider evidence with the existing scope index
rather than ranking every retained payload. Preserve history and evidence
validation. Share complete, failure-preserving backup alert evaluation
across PVE/PBS/mock polls and expose unavailable reads through the existing
deduplicated system warning; keep existing age alerts until a complete read.

Contract-Neutral: Application query and existing system-alert behaviour only; no API, schema, resource identity, agent lifecycle, entitlement or notification contract shape changes. The new warning uses the existing open system-alert type, identity and metadata contract. Semantic contracts and connected verification are included.
Change-source: pulse-maintainer
2026-10-04 15:56:57 +01:00
pulse-triage[bot] b77e6908e6 Require same-target unambiguous locks before guest-agent reads
Guest lock verification used ordinary redirect-following config reads and last-key-wins JSON decoding. Refuse redirected or ambiguous lock evidence before dispatch and after a complete guest response, without changing ordinary config redirects, pooled TLS policy or session recovery. Cover all six readers and connected cache/History continuity and fresh resumption; this source hardening does not establish native backup thaw.

Change-source: pulse-maintainer
2026-10-04 15:01:23 +01:00
pulse-triage[bot] dcde1d0157 Preserve reviewed maintenance alongside published guest memory repair
Additively compose the exact current upstream with accepted guest HTTP safeguards, runtime connection alerts, safe intake and PBS capacity evidence.

Change-source: pulse-maintainer
2026-10-04 14:06:34 +01:00
rcourtman 0060daf585 Preserve guest memory across repeated backup-lock polls
Retain the complete last-known memory value from the scoped guest snapshot only while the original successful QGA cache supports its values and age. Keep deferrals labelled previous-snapshot without renewing cache timestamps or memory History.

Exercise repeated locks after ordinary cache expiry through unified resource views, with provenance, identity, capacity and expiration controls plus fresh unlocked resumption. Native freeze-race qualification remains separate.
2026-10-04 13:15:54 +01:00
pulse-triage[bot] 7f5c8a0cf1 Restore runtime-backed TrueNAS and VMware connection alerts
Install the default monitor's snapshot callback after both platform pollers exist. Settings already observes their runtime owners; alert checks must not retain startup's nil pointers. Exercise actual saved connections, loopback failures, ledger, monitor confirmation ticks, disable and ordinary scheduled inventory recovery without changing policy or adding a surface.

Contract-Neutral: Restore existing connection-alert wiring; no route, payload, authentication, persistence or detector-policy changes.
Change-source: pulse-maintainer
2026-10-04 12:35:01 +01:00
pulse-triage[bot] cd1e3ba12f fix(proxmox): defer guest reads after uncertain HTTP failures
Preserve the actual HTTP response status rather than trusting quoted API errors. Apply the VM-wide cooldown after request timeouts and unexplained server or gateway failures, while retaining exact command-bound terminal refusals as ordinary errors. Reject ambiguous error envelopes and preserve cached guest observations alongside fresh live counters.

Change-source: pulse-maintainer
2026-10-04 11:54:46 +01:00
pulse-triage[bot] 076a4ff94f Prevent transparent replay of Proxmox guest-agent commands
Guest-agent GETs enqueue serial work: ordinary HTTP pooling, stream recovery and redirects can resend a command despite the application single-attempt guard. Use an isolated single-use HTTP/1 transport with the configured TLS, proxy, dial and timeout policy, and defer redirects without contacting their destination. Keep ordinary API recovery and current CPU/I/O counter presence/receipts unchanged. Pin wire-level lost replies, all read/redirect paths, TLS trust and connected deferred cache/History behaviour.

Change-source: pulse-maintainer
2026-10-04 10:53:56 +01:00
pulse-triage[bot] 453eccbba1 Keep platform polling independent of organisation metadata
Saved TrueNAS and VMware connections belong to validated tenant storage keys, not optional organisation metadata. A malformed org.json could prevent provider creation and block authoritative disable or removal in an unrelated tenant. Enumerate storage identities while retaining strict metadata/account readers, tenant isolation and config/credential failure handling.

Synthetic probe, polling and API controls reproduce the interval-only response family in #2382. They do not establish either reporter installation’s cause or native appliance acceptance.

Change-source: pulse-maintainer
2026-10-04 09:45:21 +01:00
pulse-triage[bot] f80f4b75ff Integrate reviewed finite-metric alert safety
Preserve the exact Core candidate and its retained focused, lifecycle, race, full-alert, monitoring and embedded-build evidence. Non-finite readings remain unknown rather than changing trusted incident state.

Change-source: pulse-maintainer
2026-10-04 09:27:53 +01:00
pulse-triage[bot] a93ab070f6 Integrate reviewed webhook receiver safety guidance
Change-source: pulse-maintainer
2026-10-04 09:04:27 +01:00
pulse-triage[bot] 1265ff1014 Keep non-finite metric evidence out of alert decisions
Treat non-finite current observations, authoritative rolling-history samples and overflowing averages as unknown. Preserve trusted incident and notification state instead of opening, recovering or poisoning JSON snapshots. Retain timestamp/duplicate authority, normal finite evaluation and explicit rule disable.

Change-source: pulse-maintainer
2026-10-04 09:02:22 +01:00
pulse-triage[bot] fcd9b65b82 Reduce temporary allocation in History reads
Count retained samples before allocating detached snapshots for guest, node, disk and storage readers. Preserve exclusive cutoffs, unordered inputs, original observations and empty arrays. Cover public series, snapshot isolation, connected chart/window consumers and same-executable allocation comparisons; no SQL, retention or source-selection change.

Change-source: pulse-maintainer
Contract-Neutral: In-memory History copy allocation only; public API/JSON shape, sample order/cutoffs, retention and observation authority are unchanged.
2026-10-04 08:18:36 +01:00
pulse-triage[bot] 94a14c78c7 Keep webhook bridge guidance from discarding later alerts
The legacy correlation header omits occurrence start, severity and group members. Show receivers the existing full-precision member data, preserve later incidents and severity updates, and test the copyable templates through the actual signed sender. Keep the freshness verifier and runtime wire format unchanged.

Contract-Neutral: Help, executable examples, tests and sender comments only; notification behaviour, schemas and webhook headers are unchanged.
Change-source: pulse-maintainer
2026-10-04 08:16:23 +01:00
pulse-triage[bot] 0ea87307e2 Keep used-memory history within its retention window
Periodic guest and node cleanup omitted raw memory-used samples from both expiry and live-entry checks. Trim them with the existing retention and backing-array policy without deleting a still-live byte series. Cover every supported History series, byte-array reclamation, concurrent cleanup and poll-to-chart/metric-window resumption. No SQL retention, source-selection or wire changes.

Change-source: pulse-maintainer
2026-10-04 06:33:37 +01:00
pulse-triage[bot] 227f4f7ef9 Require complete VM config bodies for guest-agent lock checks
Do not admit guest commands or publish their payloads from an unlocked JSON prefix when the authoritative config response is incomplete. Preserve bounded ordinary config reads, live CPU and I/O observations, cache timestamps and healthy resumption.

Change-source: pulse-maintainer
2026-10-04 04:11:34 +01:00
pulse-triage[bot] e7c2b37cf6 Keep fully cached Patrol usage in budgets and cost projections
Follow up #2118/#2350: ordinary input can be zero while cache writes or reads remain billable. Preserve cache-only successful and partial results, record them in the usage ledger, and include them in the history median and triggered-spend rate. Keep genuinely empty/expired history excluded and retain the existing five-minute prices and budget semantics.

Change-source: pulse-maintainer
2026-10-04 03:25:57 +01:00