nyyu/monitoring
1
0
Fork 0
Code Issues 1 Pull requests Projects Releases Wiki Activity

chore: upd alerts
All checks were successful
ci/woodpecker/push/woodpecker Pipeline was successful
Details

Browse source
This commit is contained in:
nyyu 2025-01-28 22:17:57 +01:00
parent 8643914e50
commit 37733b2dc8
5 changed files with 156 additions and 100 deletions
Download patch file Download diff file Expand all files Collapse all files

14
conf/prometheus/alerts/postgres-exporter.yml
View file

@ -32,7 +32,7 @@ groups:
description: "Postgresql exporter is showing errors. A query may be buggy in query.yaml\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
- alert: PostgresqlTableNotAutoVacuumed
expr: '(pg_stat_user_tables_last_autovacuum > 0) and (time() - pg_stat_user_tables_last_autovacuum) > 60 * 60 * 24 * 10'
expr: '((pg_stat_user_tables_n_tup_del + pg_stat_user_tables_n_tup_upd + pg_stat_user_tables_n_tup_hot_upd) > pg_settings_autovacuum_vacuum_threshold) and (time() - pg_stat_user_tables_last_autovacuum) > 864000'
for: 0m
labels:
severity: warning
@ -41,7 +41,7 @@ groups:
description: "Table {{ $labels.relname }} has not been auto vacuumed for 10 days\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
- alert: PostgresqlTableNotAutoAnalyzed
expr: '(pg_stat_user_tables_last_autoanalyze > 0) and (time() - pg_stat_user_tables_last_autoanalyze) > 24 * 60 * 60 * 10'
expr: '((pg_stat_user_tables_n_tup_del + pg_stat_user_tables_n_tup_upd + pg_stat_user_tables_n_tup_hot_upd) > pg_settings_autovacuum_analyze_threshold) and (time() - pg_stat_user_tables_last_autoanalyze) > 864000'
for: 0m
labels:
severity: warning
@ -53,7 +53,7 @@ groups:
expr: 'sum by (instance, job, server) (pg_stat_activity_count) > min by (instance, job, server) (pg_settings_max_connections * 0.8)'
for: 2m
labels:
severity: warning
severity: critical
annotations:
summary: Postgresql too many connections (instance {{ $labels.instance }})
description: "PostgreSQL instance has too many connections (> 80%).\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
@ -62,7 +62,7 @@ groups:
expr: 'sum by (datname) (pg_stat_activity_count{datname!~"template.*|postgres"}) < 5'
for: 2m
labels:
severity: warning
severity: critical
annotations:
summary: Postgresql not enough connections (instance {{ $labels.instance }})
description: "PostgreSQL instance should have more connections (> 5)\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
@ -86,8 +86,8 @@ groups:
description: "Ratio of transactions being aborted compared to committed is > 2 %\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
- alert: PostgresqlCommitRateLow
expr: 'rate(pg_stat_database_xact_commit[1m]) < 10'
for: 2m
expr: 'increase(pg_stat_database_xact_commit{datname!~"template.*|postgres",datid!="0"}[5m]) < 5'
for: 5m
labels:
severity: critical
annotations:
@ -155,7 +155,7 @@ groups:
severity: critical
annotations:
summary: Postgresql SSL compression active (instance {{ $labels.instance }})
description: "Database connections with SSL compression enabled. This may add significant jitter in replication delay. Replicas should turn off SSL compression via `sslcompression=0` in `recovery.conf`.\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
description: "Database allows connections with SSL compression enabled.\n VALUE = {{ $value }}\n LABELS = {{ $labels }}"
- alert: PostgresqlTooManyLocksAcquired
expr: '((sum (pg_locks_count)) / (pg_settings_max_locks_per_transaction * pg_settings_max_connections)) > 0.20'