agentic_sre_env / knowledge_base /db_pool_runbook.md
LordofMonarchs's picture
initial commit
1ee199d
|
Raw
History Blame Contribute Delete
886 Bytes

DB Pool Exhaustion Runbook

Symptoms

  • PostgreSQL connection pool at 20/20 active connections
  • Lock contention detected on auth-db
  • Cascading latency on auth-service and api-gateway

Diagnostic Steps

  1. Query pool metric on auth-db to confirm exhaustion
  2. Check lock_contention flag in pool status
  3. Inspect auth-db logs for long-running query signatures

Remediation Options

Option A — kill long-running queries:

  • remediation: kill_pid on auth-db
  • Verify pool utilisation drops below 80%

Option B — release connections:

  • remediation: update_config on auth-db with max_connections=40
  • Verify lock_contention clears

Option C — scale upstream:

  • remediation: scale_up on auth-service to reduce connection pressure

Key Metric

Pool utilisation must drop below 50% for full remediation credit. Lock contention must clear before submitting resolution.