mirror of
https://github.com/samber/awesome-prometheus-alerts.git
synced 2026-06-22 01:17:19 +08:00
50 lines
2.5 KiB
YAML
50 lines
2.5 KiB
YAML
groups:
|
|
|
|
- name: ThanosCompactor
|
|
|
|
rules:
|
|
|
|
- alert: ThanosCompactorMultipleRunning
|
|
expr: 'sum by (job) (up{job=~".*thanos-compact.*"}) > 1'
|
|
for: 5m
|
|
labels:
|
|
severity: warning
|
|
annotations:
|
|
summary: Thanos Compactor Multiple Running (instance {{ $labels.instance }})
|
|
description: "No more than one Thanos Compact instance should be running at once. There are {{$value}} instances running.\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
|
|
|
|
- alert: ThanosCompactorHalted
|
|
expr: 'thanos_compact_halted{job=~".*thanos-compact.*"} == 1'
|
|
for: 5m
|
|
labels:
|
|
severity: warning
|
|
annotations:
|
|
summary: Thanos Compactor Halted (instance {{ $labels.instance }})
|
|
description: "Thanos Compact {{$labels.job}} has failed to run and now is halted.\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
|
|
|
|
- alert: ThanosCompactorHighCompactionFailures
|
|
expr: '(sum by (job) (rate(thanos_compact_group_compactions_failures_total{job=~".*thanos-compact.*"}[5m])) / sum by (job) (rate(thanos_compact_group_compactions_total{job=~".*thanos-compact.*"}[5m])) * 100 > 5)'
|
|
for: 15m
|
|
labels:
|
|
severity: warning
|
|
annotations:
|
|
summary: Thanos Compactor High Compaction Failures (instance {{ $labels.instance }})
|
|
description: "Thanos Compact {{$labels.job}} is failing to execute {{$value | humanize}}% of compactions.\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
|
|
|
|
- alert: ThanosCompactBucketHighOperationFailures
|
|
expr: '(sum by (job) (rate(thanos_objstore_bucket_operation_failures_total{job=~".*thanos-compact.*"}[5m])) / sum by (job) (rate(thanos_objstore_bucket_operations_total{job=~".*thanos-compact.*"}[5m])) * 100 > 5)'
|
|
for: 15m
|
|
labels:
|
|
severity: warning
|
|
annotations:
|
|
summary: Thanos Compact Bucket High Operation Failures (instance {{ $labels.instance }})
|
|
description: "Thanos Compact {{$labels.job}} Bucket is failing to execute {{$value | humanize}}% of operations.\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
|
|
|
|
- alert: ThanosCompactHasNotRun
|
|
expr: '(time() - max by (job) (max_over_time(thanos_objstore_bucket_last_successful_upload_time{job=~".*thanos-compact.*"}[24h]))) / 60 / 60 > 24'
|
|
for: 0m
|
|
labels:
|
|
severity: warning
|
|
annotations:
|
|
summary: Thanos Compact Has Not Run (instance {{ $labels.instance }})
|
|
description: "Thanos Compact {{$labels.job}} has not uploaded anything for 24 hours.\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
|