We should add a Docker health check to detect cases where the keyper is hanging.
Goal
- Automatically detect when the Keyper is hanging
- Let Docker restart the container when the health check fails.
- Log restart events so we can see them in metrics
Proposed approach
- Add a health check in Dockerfile
- When the health check fails 3 times in a row, Docker marks the container as unhealthy and restarts it.
Additional information:
Instead of the ping endpoint that relies on user input, rather use a new endpoint that is not reliant on user specific configuration, but that is still configurable internally, i.e. make it a regular option in the keyper config, but don't expose it in the docker env var based setup but just set it to enabled in the config script.
We should add a Docker health check to detect cases where the keyper is hanging.
Goal
Proposed approach
Additional information:
Instead of the ping endpoint that relies on user input, rather use a new endpoint that is not reliant on user specific configuration, but that is still configurable internally, i.e. make it a regular option in the keyper config, but don't expose it in the docker env var based setup but just set it to enabled in the config script.