Skip to content

Kubernetes Probes Configuration Guide

This repository contains the Kubernetes Deployment configuration featuring Startup, Readiness, and Liveness probes for the Node.js application (sachinthokal/k8s_probes_app:v1).

alt text

🛠 Deployment Manifest (deployment.yaml)

kind: Deployment
apiVersion: apps/v1
metadata:
  name: probs-deployment
  namespace: probs-ns
spec:
  replicas: 1
  selector:
    matchLabels:
      app: probs-app
  template:
    metadata:
      labels:
        app: probs-app
    spec:
      containers:
        - name: probs-container
          image: sachinthokal/k8s_probes_app:v1
          ports:
            - containerPort: 3000

          # 1. STARTUP PROBE
          startupProbe:
            httpGet:
              path: /health/readiness
              port: 3000
            periodSeconds: 10
            failureThreshold: 30

          # 2. READINESS PROBE
          readinessProbe:
            httpGet:
              path: /health/readiness
              port: 3000
            initialDelaySeconds: 3
            periodSeconds: 3
            failureThreshold: 3
            timeoutSeconds: 2
            successThreshold: 2

          # 3. LIVENESS PROBE
          livenessProbe:
            httpGet:
              path: /health/liveness
              port: 3000
            initialDelaySeconds: 5
            periodSeconds: 5
            failureThreshold: 3
            timeoutSeconds: 2
            successThreshold: 1

🔍 How Probes Work in This Configuration

1. startupProbe (Boot-up Protection)

  • Goal: Gives the application time to perform initial startup (e.g., loading cache, DB connections).
  • Behavior: Checks http://<pod-ip>:3000/health/readiness every 10 seconds (periodSeconds).
  • Max Timeout: Allows up to 30 failures (failureThreshold). 10s × 30 = 300 seconds (5 minutes).
  • Key Rule: Both Readiness and Liveness probes remain DISABLED until the Startup probe succeeds.

2. readinessProbe (Traffic Routing)

  • Goal: Determines if the Pod is ready to accept incoming user network traffic.
  • Behavior: Starts 3 seconds after the startup probe succeeds (initialDelaySeconds), checking /health/readiness every 3 seconds (periodSeconds).
  • Timeout: Expects an HTTP response within 2 seconds (timeoutSeconds).
  • Success Rule: Requires 2 consecutive successful checks (successThreshold: 2) before marking the Pod as READY.
  • Failure Rule: If it fails 3 consecutive times (failureThreshold: 3), Kubernetes stops routing traffic to this Pod (Removes it from Service Endpoints). The container is NOT restarted.

3. livenessProbe (Container Health & Auto-Healing)

  • Goal: Detects if the application is deadlocked or frozen and needs a restart.
  • Behavior: Starts 5 seconds after the startup probe succeeds (initialDelaySeconds), checking /health/liveness every 5 seconds (periodSeconds).
  • Timeout: Expects an HTTP response within 2 seconds (timeoutSeconds).
  • Success Rule: Marks healthy on 1 successful response (successThreshold: 1).
  • Failure Rule: If it fails 3 consecutive times (failureThreshold: 3), Kubernetes considers the container dead and KILLS & RESTARTS the container automatically.

🚀 How to Deploy and Verify

1. Create Namespace & Deploy

kubectl create namespace probs-ns
kubectl apply -f deployment.yaml

2. Check Deployment Status

kubectl get pods -n probs-ns -w

3. Inspect Probe Events & Logs

# Detailed events and probe failures
kubectl describe pod <pod-name> -n probs-ns

# Stream real-time logs
kubectl logs -f deployment/probs-deployment -n probs-ns