Kubernetes Readiness Probe Health Check

Kubernetes Readiness Probe

The readiness probe tells Kubernetes whether your pod can handle traffic. Unlike liveness, a failed readiness probe removes the pod from service endpoints without restarting it.

Response Format

{
  "status": "ok",
  "dependencies": {
    "database": {
      "status": "ok",
      "responseTime": "12ms",
      "message": "Connection pool healthy"
    },
    "redis": {
      "status": "ok",
      "responseTime": "3ms",
      "message": "Cache connected"
    },
    "external-api": {
      "status": "ok",
      "responseTime": "89ms",
      "message": "Payment service responding"
    }
  }
}

Probe Configuration

readinessProbe:
  httpGet:
    path: /ready
    port: 8080
  initialDelaySeconds: 5
  periodSeconds: 5
  timeoutSeconds: 3
  failureThreshold: 2

What to Check

Check	Include?	Why
Database connection	Yes	Can't serve requests without it
Cache availability	Yes	Performance critical
External API health	Depends	If requests require it
Disk space	Sometimes	If writes are needed
Configuration loaded	Yes	Ensures correct behavior

Liveness vs Readiness

Aspect	Liveness	Readiness
Path	/healthz	/ready
Failure action	Restart pod	Remove from service
Check scope	Internal only	Dependencies too
Recovery	Automatic restart	Automatic re-add

Best Practice

Keep readiness probe timeouts short (3-5 seconds). If a dependency check takes longer, consider it failed. Use circuit breakers to prevent slow dependencies from blocking the probe.