Understanding the Docker monitoring panel and container health

Learn how to read the Admin container health dashboard, understand status indicators, and respond to unhealthy containers.

Navigate to Admin > Containers

From the Admin panel, click "Containers" in the sidebar. The monitoring panel shows the real-time health of all Docker containers that power the platform.

Understand the four container groups

Containers are organised into four groups:

• Nginx — The load balancer. Usually one or two instances. Routes incoming traffic to app containers.
• App — The Next.js application servers. Multiple instances run in parallel. These handle user requests.
• Redis — The primary Redis cache / message queue. One master, one or more replicas.
• Sentinel — The Redis Sentinel monitors that watch the Redis cluster and manage failover.

Read the health status badges

Each container shows one of three status badges:

• Healthy (green) — The container is running and passing its health check. Normal operation.
• Starting (amber) — The container has just started and is waiting for its health check to pass. This is expected briefly after a deployment.
• Unhealthy (red) — The container is running but failing its health check, or has exited unexpectedly. This requires investigation.

Read the CPU and memory usage bars

Below each container name you will see CPU % and memory usage bars. A CPU bar above 80% (shown in amber) indicates the container is under heavy load. A bar above 95% (shown in red) means the container is at risk of becoming unresponsive. If multiple app containers are all above 80%, the autoscaler should have already triggered a new instance — check the Containers log for scale-up events.

Respond to an unhealthy container

If a container turns red, first click on it to expand the logs panel. Review the last 50 log lines for error messages. Common causes include:

• Database connection timeout — check Supabase status
• Out of memory — the container may need its memory limit increased
• Application crash — check the app logs for unhandled exceptions

If you cannot resolve the issue from the logs, use the "Restart Container" button to force a clean restart.