mirror of
https://github.com/rustfs/rustfs.git
synced 2026-08-17 02:15:28 +00:00
Zhengchao An
87d47a6e5d
Add a per-backend rotation-driver matrix to docs/operations/kms-backend-security.md: who performs the rotation on each backend (RustFS for Vault KV2, Vault's Transit engine for Transit, AWS RotateKeyOnDemand for AWS, nobody for Local/Static), how periodic rotation must be scheduled on each (external scheduler for KV2 by design, Vault auto_rotate_period for Transit, AWS-native automatic rotation for AWS since RotateKeyOnDemand carries a lifetime quota), the NIST SP 800-38D 2^32 random-nonce AES-GCM wrap ceiling that Local/Static can never reset, and a pre-rotation checklist referencing the existing upgrade-ordering hard constraint. Add the KmsKeyRotationOverdue Prometheus rule on rustfs_kms_oldest_key_rotation_age_seconds (400-day conservative default, warning severity, no traffic guard because it is direct gauge state) and its runbook response procedure in docs/operations/kms-observability-runbook.md, following the existing per-alert format. Sharpen the runbook's rotation-timestamp paragraph with verified per-backend behavior: only Vault KV2 persists rotated_at (stamped in the same check-and-set write that commits the rotation), while Transit and AWS key listings always report it absent, so on those backends the gauge measures key age and does not reset on rotation. Update Threshold calibration and Coverage gaps for the new rule. Validated with promtool check rules (7 rules, SUCCESS) and scripts/check_doc_paths.sh. Part of rustfs/backlog#1636 (PR-4).
252 lines
13 KiB
YAML
252 lines
13 KiB
YAML
# 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.
|
|
|
|
# =============================================================================
|
|
# RustFS KMS backend — Prometheus alerting rules
|
|
# =============================================================================
|
|
#
|
|
# Metric source: the KMS operation-policy choke point in
|
|
# crates/kms/src/policy.rs, except KmsKeyRotationOverdue, which reads the
|
|
# label-less key-lifecycle gauge published by the deletion worker's sweep
|
|
# (crates/kms/src/deletion_worker.rs). All label values are bounded static
|
|
# strings (operation, op_class, outcome, error_class, backend, scope); key
|
|
# identifiers, key material, and tokens never appear in labels.
|
|
#
|
|
# Response procedures: docs/operations/kms-observability-runbook.md
|
|
#
|
|
# IMPORTANT — threshold status: every numeric threshold below is a
|
|
# conservative default chosen without a production baseline. Calibrate against
|
|
# a staging baseline before relying on these alerts for paging, and prefer
|
|
# loosening over tightening until the baseline exists. Formal SLO targets are
|
|
# deliberately not encoded here (see rustfs/backlog#1584).
|
|
#
|
|
# NOTE: prometheus.yml loads /etc/prometheus/rules/*.yml — keep the .yml
|
|
# extension or the file is silently ignored by the docker-compose stack.
|
|
#
|
|
# Validate: promtool check rules rustfs-kms-alerts.yml
|
|
# =============================================================================
|
|
|
|
groups:
|
|
# ==========================================================================
|
|
# Critical alerts — immediate action required
|
|
# ==========================================================================
|
|
- name: rustfs-kms-critical
|
|
interval: 30s
|
|
rules:
|
|
# ------------------------------------------------------------------
|
|
# 1. KmsBackendFatalErrors
|
|
# Any attempt failure classified as fatal (non-retryable): auth
|
|
# or permission errors, malformed requests, missing keys. The
|
|
# policy never retries these, so even a low rate means real
|
|
# operations are failing right now.
|
|
# ------------------------------------------------------------------
|
|
- alert: KmsBackendFatalErrors
|
|
expr: |
|
|
sum by (operation) (rate(rustfs_kms_backend_attempt_failures_total{error_class="fatal"}[5m])) > 0
|
|
for: 5m
|
|
labels:
|
|
severity: critical
|
|
component: kms
|
|
annotations:
|
|
summary: "KMS backend fatal errors on operation {{ $labels.operation }}"
|
|
description: >-
|
|
Attempt failures classified as fatal are occurring at
|
|
{{ $value | printf "%.3f" }}/s on operation
|
|
{{ $labels.operation }}. Fatal failures are not retried:
|
|
each one is a KMS backend call that failed permanently
|
|
(authentication, permissions, malformed request, or a
|
|
missing key/version).
|
|
runbook_url: "https://github.com/rustfs/rustfs/blob/main/docs/operations/kms-observability-runbook.md#kmsbackendfatalerrors"
|
|
|
|
# ------------------------------------------------------------------
|
|
# 2. KmsBackendHighErrorRate
|
|
# Sustained share of operations terminating without success
|
|
# (fatal, budget/deadline exhaustion, admission backpressure,
|
|
# or an open circuit). The cancelled outcome is excluded because
|
|
# shutdowns legitimately produce it.
|
|
# The traffic guard keeps a single failure on a near-idle
|
|
# cluster from firing the alert.
|
|
# Threshold: 5% for 10m — conservative default, calibrate
|
|
# against a staging baseline.
|
|
# ------------------------------------------------------------------
|
|
- alert: KmsBackendHighErrorRate
|
|
expr: |
|
|
(
|
|
sum(rate(rustfs_kms_backend_operations_total{outcome!~"success|cancelled"}[5m]))
|
|
/
|
|
clamp_min(sum(rate(rustfs_kms_backend_operations_total[5m])), 1e-9)
|
|
) > 0.05
|
|
and
|
|
sum(rate(rustfs_kms_backend_operations_total[5m])) > 0.02
|
|
for: 10m
|
|
labels:
|
|
severity: critical
|
|
component: kms
|
|
annotations:
|
|
summary: "KMS backend non-success ratio above 5% for 10m"
|
|
description: >-
|
|
{{ $value | humanizePercentage }} of KMS backend operations
|
|
are terminating in fatal, budget_exhausted,
|
|
deadline_exceeded, backpressure_timeout,
|
|
backpressure_rejected, or circuit_open. Object encryption
|
|
and decryption paths depending on the KMS are degraded or
|
|
failing.
|
|
runbook_url: "https://github.com/rustfs/rustfs/blob/main/docs/operations/kms-observability-runbook.md#kmsbackendhigherrorrate"
|
|
|
|
# ==========================================================================
|
|
# Warning alerts — investigation needed
|
|
# ==========================================================================
|
|
- name: rustfs-kms-warning
|
|
interval: 30s
|
|
rules:
|
|
# ------------------------------------------------------------------
|
|
# 3. KmsBackendP99LatencyHigh
|
|
# p99 wall-clock duration of whole operations (attempts plus
|
|
# backoff) is sustained above 2 seconds. Because the histogram
|
|
# includes retries, a high p99 usually means the retry policy
|
|
# is absorbing backend failures, not that every call is slow.
|
|
# Threshold: 2s for 10m — conservative default, calibrate
|
|
# against a staging baseline.
|
|
# ------------------------------------------------------------------
|
|
- alert: KmsBackendP99LatencyHigh
|
|
expr: |
|
|
histogram_quantile(0.99,
|
|
sum by (le) (rate(rustfs_kms_backend_operation_duration_seconds_bucket[5m]))
|
|
) > 2
|
|
for: 10m
|
|
labels:
|
|
severity: warning
|
|
component: kms
|
|
annotations:
|
|
summary: "KMS backend operation p99 latency above 2s for 10m"
|
|
description: >-
|
|
The 99th-percentile KMS backend operation duration is
|
|
{{ $value | humanizeDuration }}, including retries and
|
|
backoff. Encryption and decryption latency is leaking into
|
|
S3 request latency.
|
|
runbook_url: "https://github.com/rustfs/rustfs/blob/main/docs/operations/kms-observability-runbook.md#kmsbackendp99latencyhigh"
|
|
|
|
# ------------------------------------------------------------------
|
|
# 4. KmsBackendAttemptFailureSpike
|
|
# Aggregate attempt-failure rate (all error classes) sustained
|
|
# above an absolute floor. An absolute threshold is used instead
|
|
# of an offset-1d baseline ratio because fresh deployments have
|
|
# no baseline and an empty offset vector would keep a ratio
|
|
# alert from ever firing; switch to a baseline-relative form
|
|
# (see rustfs-get-optimization-alerts.yaml for the pattern)
|
|
# once a stable staging baseline exists.
|
|
# Threshold: 0.5/s for 10m — conservative default, calibrate
|
|
# against a staging baseline.
|
|
# ------------------------------------------------------------------
|
|
- alert: KmsBackendAttemptFailureSpike
|
|
expr: |
|
|
sum(rate(rustfs_kms_backend_attempt_failures_total[5m])) > 0.5
|
|
for: 10m
|
|
labels:
|
|
severity: warning
|
|
component: kms
|
|
annotations:
|
|
summary: "KMS backend attempt failures above 0.5/s for 10m"
|
|
description: >-
|
|
KMS backend attempts are failing at
|
|
{{ $value | printf "%.2f" }}/s across all error classes.
|
|
The retry policy may still be masking these from callers —
|
|
check the error-class breakdown before it stops absorbing
|
|
them.
|
|
runbook_url: "https://github.com/rustfs/rustfs/blob/main/docs/operations/kms-observability-runbook.md#kmsbackendattemptfailurespike"
|
|
|
|
# ------------------------------------------------------------------
|
|
# 5. KmsBackendRetryBudgetExhausted
|
|
# Operations are running out of retry budget (budget_exhausted)
|
|
# or operation deadline (deadline_exceeded). These surface to
|
|
# callers as failed KMS operations even though every individual
|
|
# failure was retryable — the backend is unhealthy for longer
|
|
# than the policy can bridge.
|
|
# Threshold: 0.05/s for 10m — conservative default, calibrate
|
|
# against a staging baseline.
|
|
# ------------------------------------------------------------------
|
|
- alert: KmsBackendRetryBudgetExhausted
|
|
expr: |
|
|
sum by (outcome) (rate(rustfs_kms_backend_operations_total{outcome=~"budget_exhausted|deadline_exceeded"}[5m])) > 0.05
|
|
for: 10m
|
|
labels:
|
|
severity: warning
|
|
component: kms
|
|
annotations:
|
|
summary: "KMS backend operations exhausting retry budget ({{ $labels.outcome }})"
|
|
description: >-
|
|
KMS backend operations are terminating as
|
|
{{ $labels.outcome }} at {{ $value | printf "%.3f" }}/s.
|
|
Retryable failures are outlasting the retry budget, so
|
|
callers are seeing hard failures.
|
|
runbook_url: "https://github.com/rustfs/rustfs/blob/main/docs/operations/kms-observability-runbook.md#kmsbackendretrybudgetexhausted"
|
|
|
|
# ------------------------------------------------------------------
|
|
# 6. KmsBackendCircuitOpen
|
|
# Direct circuit-state signal, independent of operation traffic.
|
|
# A transient open can recover on its first half-open probe; alert
|
|
# only when the circuit remains open or half-open for one minute.
|
|
# ------------------------------------------------------------------
|
|
- alert: KmsBackendCircuitOpen
|
|
expr: |
|
|
rustfs_kms_backend_circuit_open > 0
|
|
for: 1m
|
|
labels:
|
|
severity: warning
|
|
component: kms
|
|
annotations:
|
|
summary: "KMS backend circuit open ({{ $labels.backend }}/{{ $labels.scope }})"
|
|
description: >-
|
|
The KMS backend circuit for {{ $labels.backend }} scope
|
|
{{ $labels.scope }} has remained open or half-open for one
|
|
minute. Operations in this scope can terminate as
|
|
circuit_open until the half-open probe succeeds or returns
|
|
a non-retryable failure.
|
|
runbook_url: "https://github.com/rustfs/rustfs/blob/main/docs/operations/kms-observability-runbook.md#kmsbackendcircuitopen"
|
|
|
|
# ------------------------------------------------------------------
|
|
# 7. KmsKeyRotationOverdue
|
|
# The least recently rotated usable key has gone more than 400
|
|
# days without a rotation (measured from creation for keys with
|
|
# no recorded rotation). Direct gauge state published by the
|
|
# deletion worker's sweep, so no traffic guard applies; the
|
|
# one-hour hold only bridges scrape gaps. The worker runs only
|
|
# on backends with the schedule_deletion capability, so on the
|
|
# Static backend the series never exists and this alert cannot
|
|
# fire — that backend cannot rotate either; see the rotation
|
|
# driver matrix in docs/operations/kms-backend-security.md.
|
|
# Threshold: 400 days — conservative default sitting above a
|
|
# one-year rotation policy. Align it with the rotation period
|
|
# your compliance policy requires, and with
|
|
# RUSTFS_KMS_ROTATION_MAX_AGE_SECS so the per-key rotation_due
|
|
# verdict and this aggregate alert agree.
|
|
# ------------------------------------------------------------------
|
|
- alert: KmsKeyRotationOverdue
|
|
expr: |
|
|
rustfs_kms_oldest_key_rotation_age_seconds > (400 * 86400)
|
|
for: 1h
|
|
labels:
|
|
severity: warning
|
|
component: kms
|
|
annotations:
|
|
summary: "Oldest KMS key unrotated for more than 400 days"
|
|
description: >-
|
|
The least recently rotated usable KMS key was last rotated
|
|
{{ $value | humanizeDuration }} ago (measured from creation
|
|
for keys with no recorded rotation). List keys through the
|
|
admin API and read rotation_due / rotation_due_reason for
|
|
the per-key verdict; an "unsupported" reason means the
|
|
backend cannot rotate at all.
|
|
runbook_url: "https://github.com/rustfs/rustfs/blob/main/docs/operations/kms-observability-runbook.md#kmskeyrotationoverdue"
|