A Docker Healthcheck Must Test the Dependency You Actually Need

A process can be alive while the service contract its neighbors depend on is broken.

I keep this as a field note because the failure mode is easy to misclassify: health is defined as process existence instead of useful readiness. The useful move is to identify the boundary first, then change only the layer that owns it.

What I model

The system is easier to debug when intent, observation and transport are not collapsed into one state. For this case, my rule is simple: Health should measure the contract consumers rely on.

Implementation pattern

Use a fast side-effect-free readiness probe against the actual interface or dependency required by downstream services.

I prefer a small explicit contract over a clever implicit one. That gives logs, tests and dashboards something concrete to verify and keeps unrelated layers from compensating for each other.

What I verify

• the check fails on real dependency loss
• the check remains cheap
• restart policy does not hide persistent failure

Failure handling

When one of those checks fails, I preserve the failing evidence before restarting or changing configuration. The first broken contract determines the next investigation. That keeps troubleshooting causal instead of turning it into a sequence of guesses.

What I keep

Health should measure the contract consumers rely on. The specific tools can change, but that ownership boundary remains useful across firmware, networks and infrastructure.