DOCKER HEALTHCHECK

The HEALTHCHECK instruction in a Dockerfile is a built-in mechanism to monitor container health. Let me break it down:

What it does:

1. Defines a health check command that Docker will execute periodically
2. Reports container health status (starting, healthy, or unhealthy)
3. Helps detect hung/unresponsive containers even if the process is still running

From your Dockerfile example:

HEALTHCHECK CMD curl --fail http://localhost:8501/_stcore/health

How it works:

1. Check Command (curl --fail ...):

   • Tries to access Streamlit's health endpoint
   • --fail makes curl return non-zero exit code on HTTP errors
   • Only returns success (0) if HTTP status is 200-399

2. Default Timing (if not specified):

   • First check: 30s after container starts
   • Subsequent checks: Every 30 seconds
   • Timeout: 30 seconds per check
   • Retries: 3 consecutive failures mark container as unhealthy

Why it's useful for your Streamlit app:

1. Automatic recovery: Orchestrators (like Docker Compose/Kubernetes) can automatically restart unhealthy containers
2. Visibility: See container health status in docker ps CONTAINER ID STATUS PORTS NAMES a1b2c3d4e5f6 Up 2 minutes (healthy) 0.0.0.0:8501->8501/tcp pg-viewer
3. Prevent routing traffic to broken instances in load-balanced setups

How to customize (optional):

HEALTHCHECK --interval=1m --timeout=10s --start-period=30s --retries=3 \
    CMD curl --fail http://localhost:8501/_stcore/health

• --interval: Time between checks (default: 30s)
• --timeout: Time before considering check failed (default: 30s)
• --start-period: Initial container startup time (default: 0s)
• --retries: Consecutive failures needed to mark unhealthy (default: 3)

For your specific use case:

• The check verifies that:
  1. Streamlit server is responding
  2. Internal health endpoint is working
  3. Container can handle HTTP requests

To test health status:

docker inspect --format='{{json .State.Health}}' <container-name>

This helps ensure your database viewer remains operational and automatically flags issues if: - Streamlit process dies but container remains running - Application enters a zombie state - Server becomes unresponsive due to errors