Today i had a curious case of a pod in openshift with problems to start. The pod contained a simple enough springboot rest web service. The logs looked all fine, But nevertheless openshift would restart the service 4 times and give up with a fairly meaningless error message.
Turns the dependency for the actuator was missing, so openshift could not reach the two endpoints /actuator/health and /actuator/info. These endpoints were configured in the openshift deployment config for the liveness probe and the readiness probe. So openshift could not know the spring app was started successfully and gave up after 4 tries.
I will add two tests for those endpoints.