Ask Claude to Add Health Check Endpoints That Monitoring Tools Can Hit

Your app might boot fine but break five minutes later when Redis goes down or the disk fills up. Health check endpoints give monitoring tools a way to ask "are you still healthy?" at any time.

"Add health check endpoints to this project — /health for a quick 
liveness check and /ready for a full readiness check of all dependencies"

Claude reads your project to understand which services you depend on, then generates endpoints that check them:

// GET /health → fast liveness check
{ "status": "ok", "timestamp": "2024-03-15T14:30:00Z" }

// GET /ready → full dependency check
{
  "status": "healthy",
  "checks": {
    "database": { "status": "up", "latency_ms": 2 },
    "redis": { "status": "up", "latency_ms": 1 },
    "storage": { "status": "up", "free_gb": 45.2 },
    "queue": { "status": "up", "pending_jobs": 12 }
  }
}

You can customize what gets checked:

# Check external APIs your app depends on
"Add checks for the payment gateway, email service, and CDN availability"

# Add degraded states
"Return 'degraded' if non-critical services are down, 'unhealthy' 
only if the app can't serve requests at all"

# Add version info for debugging
"Include the app version, deploy timestamp, and git commit hash 
in the health response"

# Kubernetes-specific
"Add separate /healthz (liveness) and /readyz (readiness) endpoints 
following Kubernetes conventions"

Wire the endpoints to your infrastructure:

• Load balancers hit /health to route traffic away from unhealthy instances
• Kubernetes uses /readyz to decide when to send traffic after a deploy
• Uptime monitors poll /ready and alert when dependencies fail
• Dashboards display the detailed check results for the on-call team

You can't fix what you can't see — health check endpoints make your app's internal state visible to the systems that keep it running.

via Claude Code