Health Checks: Making Sure Your Self-Hosted Apps Stay Alive

What Are Health Checks?

A health check is a periodic test that verifies an application is working correctly. If the check fails, automated systems can restart the application or remove it from load balancing.

Types of Health Checks

Liveness Check

"Is the process alive?" The most basic check — is the container running? If not, restart it.

Readiness Check

"Can the app handle requests?" The app might be running but not ready — still loading data, waiting for database connection, etc. Don't send traffic until it's ready.

Startup Check

"Has the app finished starting?" Some apps take a while to start. The startup check prevents premature liveness failures during initialization.

Deep Health Check

"Are all dependencies working?" Check database connectivity, cache availability, external API reachability.

Implementing Health Checks

HTTP Endpoint

The most common approach. Add a /health or /healthz endpoint that returns:

A good health response includes:

{

"status": "healthy",

"database": "connected",

"cache": "connected",

"uptime": "3d 4h 12m"

}

TCP Check

For non-HTTP services (databases, Redis), check if the port accepts connections.

Command Check

Run a command inside the container. Exit code 0 = healthy, non-zero = unhealthy.

Container Health Checks

Docker and Podman support HEALTHCHECK in Dockerfiles:

HEALTHCHECK --interval=30s --timeout=5s --retries=3 CMD curl -f http://localhost:8080/health || exit 1

If the health check fails 3 times, the container is marked unhealthy and can be automatically restarted.

Best Practices

Check Intervals

Timeout

Health checks should be fast. If a check takes more than 5 seconds, something is wrong.

Don't Check External Dependencies in Liveness

If your database is down, restarting the app won't fix it. Use deep health checks for dependencies, but don't trigger restarts based on them.

Grace Period

Give apps time to start before health checking. A 30-60 second startup grace period prevents unnecessary restarts.

TinyPod Health Checks

TinyPod automatically monitors container health. If a container becomes unresponsive, it's restarted automatically. You can configure custom health check endpoints in the dashboard for deeper monitoring.