Ilia Denisov
e922f5f3b9
CI / changes (pull_request) Successful in 4s
CI / unit (pull_request) Successful in 12s
CI / integration (pull_request) Successful in 24s
CI / ui (pull_request) Successful in 1m9s
CI / conformance (pull_request) Successful in 10s
CI / gate (pull_request) Successful in 0s
CI / deploy (pull_request) Successful in 1m41s
Add pgBackRest-based PITR for the production database, shipped disarmed (archive_mode and the base-backup timer gated off) so it stays inert until the operator arms it before real payments — an un-armed deploy cannot pile WAL onto the disk. - Postgres now builds from a thin image carrying pgBackRest (archive_command runs in-process); the prod overlay wires an encrypted (AES-256-CBC), path-style S3 repository with 30-day retention and a daily full base-backup systemd timer. - Repository config/secrets flow through the PROD_PGBACKREST_* Gitea set and write-prod-env.sh; prod-rollback re-renders the same env so a rollback cannot disarm archiving. - Grafana alerts watch pg_stat_archiver (failing / stalled), absent-safe on the test contour, which never archives. - Fix the pre-migration pg_dump to snapshot the whole database (was backend-only, silently excluding payments). - Document the PITR runbook, the arming sequence and the restore drill in deploy/README.md; record the measured cost/perf assessment.
274 lines
9.0 KiB
YAML
274 lines
9.0 KiB
YAML
# Grafana provisioned alert rules for the Scrabble contour. Each rule is a Prometheus
|
|
# instant query (refId A) fed into a threshold expression (refId C). noDataState/execErrState
|
|
# are OK so an absent metric never raises a false alert — notably cert-expiry, whose blackbox
|
|
# probe has no target on the test contour (caddy is HTTP-only there). Metric names are the
|
|
# real ones Prometheus exposes (edge_request_* from the gateway via the collector,
|
|
# node_*/pg_*/probe_ssl_* from the exporters). All route to the ops-email contact point.
|
|
apiVersion: 1
|
|
groups:
|
|
- orgId: 1
|
|
name: scrabble-service
|
|
folder: Alerts
|
|
interval: 1m
|
|
rules:
|
|
- uid: svc_target_down
|
|
title: Scrape target down
|
|
condition: C
|
|
for: 3m
|
|
noDataState: OK
|
|
execErrState: OK
|
|
data:
|
|
- refId: A
|
|
relativeTimeRange: { from: 300, to: 0 }
|
|
datasourceUid: prometheus
|
|
model: { refId: A, expr: up, instant: true }
|
|
- refId: C
|
|
datasourceUid: __expr__
|
|
model:
|
|
refId: C
|
|
type: threshold
|
|
expression: A
|
|
conditions:
|
|
- evaluator: { type: lt, params: [1] }
|
|
labels: { severity: critical }
|
|
annotations: { summary: 'A Prometheus scrape target is down (up < 1).' }
|
|
|
|
- uid: edge_error_rate
|
|
title: Gateway internal-error rate high
|
|
condition: C
|
|
for: 10m
|
|
noDataState: OK
|
|
execErrState: OK
|
|
data:
|
|
- refId: A
|
|
relativeTimeRange: { from: 600, to: 0 }
|
|
datasourceUid: prometheus
|
|
model:
|
|
refId: A
|
|
expr: sum by (service_name) (rate(edge_request_duration_count{result="internal"}[5m]))
|
|
instant: true
|
|
- refId: C
|
|
datasourceUid: __expr__
|
|
model:
|
|
refId: C
|
|
type: threshold
|
|
expression: A
|
|
conditions:
|
|
- evaluator: { type: gt, params: [0.05] }
|
|
labels: { severity: warning }
|
|
annotations: { summary: 'Sustained internal (5xx-equivalent) errors at the edge.' }
|
|
|
|
- uid: edge_latency_p99
|
|
title: Gateway request latency p99 high
|
|
condition: C
|
|
for: 10m
|
|
noDataState: OK
|
|
execErrState: OK
|
|
data:
|
|
- refId: A
|
|
relativeTimeRange: { from: 600, to: 0 }
|
|
datasourceUid: prometheus
|
|
model:
|
|
refId: A
|
|
expr: histogram_quantile(0.99, sum by (le, service_name) (rate(edge_request_duration_bucket[5m])))
|
|
instant: true
|
|
- refId: C
|
|
datasourceUid: __expr__
|
|
model:
|
|
refId: C
|
|
type: threshold
|
|
expression: A
|
|
conditions:
|
|
- evaluator: { type: gt, params: [1] }
|
|
labels: { severity: warning }
|
|
annotations: { summary: 'Edge request p99 latency above 1s.' }
|
|
|
|
- uid: tls_cert_expiry
|
|
title: TLS certificate nearing expiry
|
|
condition: C
|
|
for: 15m
|
|
noDataState: OK
|
|
execErrState: OK
|
|
data:
|
|
- refId: A
|
|
relativeTimeRange: { from: 300, to: 0 }
|
|
datasourceUid: prometheus
|
|
model:
|
|
refId: A
|
|
expr: (probe_ssl_earliest_cert_expiry - time()) / 86400
|
|
instant: true
|
|
- refId: C
|
|
datasourceUid: __expr__
|
|
model:
|
|
refId: C
|
|
type: threshold
|
|
expression: A
|
|
conditions:
|
|
- evaluator: { type: lt, params: [20] }
|
|
labels: { severity: critical }
|
|
annotations: { summary: 'Edge TLS cert has under 20 days left — Caddy ACME renewal may have failed.' }
|
|
|
|
- orgId: 1
|
|
name: scrabble-host
|
|
folder: Alerts
|
|
interval: 1m
|
|
rules:
|
|
- uid: host_mem_low
|
|
title: Host memory low
|
|
condition: C
|
|
for: 5m
|
|
noDataState: OK
|
|
execErrState: OK
|
|
data:
|
|
- refId: A
|
|
relativeTimeRange: { from: 300, to: 0 }
|
|
datasourceUid: prometheus
|
|
model:
|
|
refId: A
|
|
expr: node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes
|
|
instant: true
|
|
- refId: C
|
|
datasourceUid: __expr__
|
|
model:
|
|
refId: C
|
|
type: threshold
|
|
expression: A
|
|
conditions:
|
|
- evaluator: { type: lt, params: [0.1] }
|
|
labels: { severity: critical }
|
|
annotations: { summary: 'Under 10% host memory available.' }
|
|
|
|
- uid: host_disk_low
|
|
title: Host disk low
|
|
condition: C
|
|
for: 10m
|
|
noDataState: OK
|
|
execErrState: OK
|
|
data:
|
|
- refId: A
|
|
relativeTimeRange: { from: 300, to: 0 }
|
|
datasourceUid: prometheus
|
|
model:
|
|
refId: A
|
|
expr: min(node_filesystem_avail_bytes / node_filesystem_size_bytes)
|
|
instant: true
|
|
- refId: C
|
|
datasourceUid: __expr__
|
|
model:
|
|
refId: C
|
|
type: threshold
|
|
expression: A
|
|
conditions:
|
|
- evaluator: { type: lt, params: [0.1] }
|
|
labels: { severity: critical }
|
|
annotations: { summary: 'A host filesystem is under 10% free.' }
|
|
|
|
- uid: host_cpu_high
|
|
title: Host CPU saturated
|
|
condition: C
|
|
for: 10m
|
|
noDataState: OK
|
|
execErrState: OK
|
|
data:
|
|
- refId: A
|
|
relativeTimeRange: { from: 600, to: 0 }
|
|
datasourceUid: prometheus
|
|
model:
|
|
refId: A
|
|
expr: 1 - avg(rate(node_cpu_seconds_total{mode="idle"}[5m]))
|
|
instant: true
|
|
- refId: C
|
|
datasourceUid: __expr__
|
|
model:
|
|
refId: C
|
|
type: threshold
|
|
expression: A
|
|
conditions:
|
|
- evaluator: { type: gt, params: [0.9] }
|
|
labels: { severity: warning }
|
|
annotations: { summary: 'Host CPU above 90% for 10 minutes.' }
|
|
|
|
- uid: pg_connections_high
|
|
title: Postgres connections high
|
|
condition: C
|
|
for: 5m
|
|
noDataState: OK
|
|
execErrState: OK
|
|
data:
|
|
- refId: A
|
|
relativeTimeRange: { from: 300, to: 0 }
|
|
datasourceUid: prometheus
|
|
model:
|
|
refId: A
|
|
expr: sum(pg_stat_activity_count) / max(pg_settings_max_connections)
|
|
instant: true
|
|
- refId: C
|
|
datasourceUid: __expr__
|
|
model:
|
|
refId: C
|
|
type: threshold
|
|
expression: A
|
|
conditions:
|
|
- evaluator: { type: gt, params: [0.8] }
|
|
labels: { severity: warning }
|
|
annotations: { summary: 'Postgres using over 80% of max_connections.' }
|
|
|
|
# Continuous WAL archiving health (pgBackRest -> S3, point-in-time recovery). Both rules
|
|
# read the postgres_exporter's built-in pg_stat_archiver metrics and are absent/NaN-safe:
|
|
# on the test contour (and any host before archiving is armed) archive_mode is off, so
|
|
# failed_count stays 0 and last_archive_age is NaN — neither threshold trips.
|
|
- orgId: 1
|
|
name: scrabble-backup
|
|
folder: Alerts
|
|
interval: 1m
|
|
rules:
|
|
- uid: pg_archive_failing
|
|
title: WAL archiving failing
|
|
condition: C
|
|
for: 15m
|
|
noDataState: OK
|
|
execErrState: OK
|
|
data:
|
|
- refId: A
|
|
relativeTimeRange: { from: 900, to: 0 }
|
|
datasourceUid: prometheus
|
|
model:
|
|
refId: A
|
|
expr: increase(pg_stat_archiver_failed_count[15m])
|
|
instant: true
|
|
- refId: C
|
|
datasourceUid: __expr__
|
|
model:
|
|
refId: C
|
|
type: threshold
|
|
expression: A
|
|
conditions:
|
|
- evaluator: { type: gt, params: [0] }
|
|
labels: { severity: critical }
|
|
annotations: { summary: 'pgBackRest archive_command is failing — PITR is degrading and pg_wal can fill the disk. Run `docker exec scrabble-postgres pgbackrest --stanza=scrabble check`.' }
|
|
|
|
- uid: pg_archive_stalled
|
|
title: WAL archiving stalled
|
|
condition: C
|
|
for: 15m
|
|
noDataState: OK
|
|
execErrState: OK
|
|
data:
|
|
- refId: A
|
|
relativeTimeRange: { from: 300, to: 0 }
|
|
datasourceUid: prometheus
|
|
model:
|
|
refId: A
|
|
expr: pg_stat_archiver_last_archive_age
|
|
instant: true
|
|
- refId: C
|
|
datasourceUid: __expr__
|
|
model:
|
|
refId: C
|
|
type: threshold
|
|
expression: A
|
|
conditions:
|
|
- evaluator: { type: gt, params: [1800] }
|
|
labels: { severity: critical }
|
|
annotations: { summary: 'No WAL segment archived for over 30 minutes (archive_timeout is 5m) — archiving is stalled; the PITR recovery point is frozen and pg_wal may grow.' }
|