CI Waiting for Service/Healthcheck Timed Out

By Kaveh Alemi·Latchkey

The job gave up waiting for a dependency (database, queue, app container) to pass its healthcheck. Service-not-ready is a mechanical, transient condition: the service often just needs a moment longer.

What this error means

A wait/healthcheck step fails with timed out waiting for ... or a connection refused on a service port. A re-run usually passes once the service starts in time.

wait-for-it.sh: timeout occurred after waiting 30 seconds for postgres:5432
##[error] Service container 'db' failed to become healthy.

Common causes

The service started slower than the wait allowed

A database or app container can take longer than the fixed wait on a cold or busy runner.

A too-short or missing readiness wait

Connecting before the service is ready (or with too short a timeout) fails intermittently.

How to fix it

Wait for actual readiness with a generous timeout

Poll the real health endpoint/port instead of a fixed sleep.

for i in $(seq 1 60); do
  pg_isready -h localhost -p 5432 && break
  sleep 2
done

Use built-in healthchecks

1. Define a service health-cmd/healthcheck so CI waits on real readiness.
2. Increase the readiness timeout for slow-starting services.
3. Start the service early so it is warm by the time tests run.

How to prevent it

• Poll readiness, never a fixed sleep.
• Give slow services a generous readiness window.
• Service-not-ready is mechanical and transient: managed runners detect it, automatically retry, and offer warm capacity so the dependency is more likely ready, so a one-off blip does not fail the build.

Frequently asked questions

Related guides

References

• GitHub Actions documentation