Kubernetes Readiness and Liveness Probe demo application.

├── cmd
│   └── application
│       └── main.go
├── Dockerfile
├── LICENSE
├── Makefile
├── manifests
│   ├── deployment.yaml
│   └── service.yaml
└── README.md - This file.

An application that receives a signal (e.g., when Kubernetes / OpenShift) wants to terminate the Pod, instantly exits / terminates because of the signal.

Why is that a problem?

It takes a short amount of time till this change to a Pod is propagated across the whole Kubernetes / OpenShift cluster. This means that in the "worst" case, user requests are going against an application Pod which is currently being terminated.

"Is the application ready to accept and serve requests?"

Examples:

• Listener started and ready to process requests.
  • Some applications open the listener before everything is initialized.
• Application is able to process requests.

"Is the application able to handle requests?"

Examples:

• Not in a Deadlock.
  • Listener still working.
• Application logic still working.

Application should gracefully handle, e.g., a database which is not available right now. An application should not panick and exit because of that.