Module 7: Self-Healing & Probes

🏗️ 1. The Manual Task: Custom Watchdog Scripts

In a manual infrastructure, sysadmins used to write scripts to check if their application was still running.

The Manual Steps:

1. Write Health Checks: Bash scripts that “ping” an endpoint (e.g., curl localhost:8080/health).
2. Monitor Health: Scripts that run on a cron schedule and run systemctl restart if the app fails.

---

🩺 2. The Kubernetes Abstraction: Probes & ReplicaSets

Kubernetes provides built-in self-healing mechanisms at both the Container and Pod level.

• Liveness Probes: “Is the app deadlocked?” If yes, Kubernetes kills and restarts the container.
• Readiness Probes: “Is the app ready for traffic?” If no, the Service stops sending you requests.
• ReplicaSet Controller: If a whole node dies, it automatically restarts the Pods on a healthy node.