30_Readiness_Probe.md

What is a Readiness Probe?

• A Readiness Probe checks whether a container is ready to receive traffic.
• If the probe fails, the pod is removed/stopped from Service endpoints, but not restarted.

Five Characteristics of Readiness Probe

1. Controls traffic flow to the pod

• Pod is added to or removed from Service load balancer depending on probe result.

2. Prevents sending traffic to unready applications

• Useful during:
  • Long startup time
  • App warming
  • Database connection wait
  • Caches loading

3. Does NOT restart containers

• Failure does not kill the pod (unlike Liveness Probe).

4. Supports three probe types

• httpGet
• exec
• tcpSocket

5. Fully configurable with timing parameters

• initialDelaySeconds
• periodSeconds
• timeoutSeconds
• failureThreshold
• successThreshold (often >1 for readiness)

1. HTTP GET Readiness Probe

apiVersion: v1
kind: Pod
metadata:
  name: readiness-http
spec:
  containers:
  - name: myapp
    image: nginx
    readinessProbe:
      httpGet:
        path: /ready
        port: 8080
      initialDelaySeconds: 5       # wait before first probe
      periodSeconds: 5             # how often to check
      timeoutSeconds: 2
      successThreshold: 1
      failureThreshold: 3          # after 3 failures, mark as NotReady

2. Exec Readiness Probe

apiVersion: v1
kind: Pod
metadata:
  name: readiness-exec
spec:
  containers:
  - name: myapp
    image: busybox
    args: ["sh", "-c", "sleep 3600"]
    readinessProbe:
      exec:
        command:
        - cat
        - /tmp/is_ready
      initialDelaySeconds: 3
      periodSeconds: 5

3. TCP Readiness Probe

apiVersion: v1
kind: Pod
metadata:
  name: readiness-tcp
spec:
  containers:
  - name: myapp
    image: nginx
    readinessProbe:
      tcpSocket:
        port: 80
      initialDelaySeconds: 5
      periodSeconds: 5