Skip to content
All outages

Incident · Easy · Databases · URL shortener

Healthy, but can't connect

  • SEV2
  • TWO SHORT OUTAGES
  • SMALL TEAM
90s

The situation

Requests to your app start failing twice in one day, for about 7 and then 13 minutes. Database CPU is low. Health checks on your workers are green. When you try to open a database console to look around, you are refused as well.

The evidence

Database connections vs CPUillustrative
  • connections
  • CPU %
Logs and signalsReconstructed
10:02db: FATAL: too many connections
10:02worker health check: 200 OK (already holds a connection)
10:03new worker: cannot connect
10:03admin console: connection refused
-- max_connections: same value since launch

Your call

What is the problem?

Charts and logs are reconstructions. Aim for 90 seconds; going over costs nothing.

Last updated: Official postmortem