Below you can find an excerpt from a toml configuration file for a service with healthcheck enabled:

port = 3000
initialDelaySeconds = 30
probePeriodSeconds = 60

Healthcheck behaviour

The healthcheck endpoint must be available at /healthz on the port specified in healthCheck.port. It is not necessary to publish the port in order to enable healthchecks on it. However, if the container image itself exposes multiple ports, only one of them can be used for healthchecks.

The endpoint must return 200 status code within 5 seconds for the service to be considered healthy. If the service fails to respond correctly 3 times in a row it will be considered not healthy and will be restarted with an exponential backoff delay (10s, 20s, 40s, …), that is capped at five minutes.

With healthCheck.initialDelaySeconds you can specify the number of seconds to wait before sending the first healthcheck request, this is useful if the service needs some time to start up. The default value is 30 seconds.

With healthCheck.probePeriodSeconds you can specify the number of seconds to wait between healthcheck requests. The default value is 60 seconds.