Health and resource monitoring

DockDash monitors whether services are available and, for containers, how many resources they consume. Current readings and historical samples appear together in the service drawer.

Service status

The health check depends on where a service came from:

• Docker: the container is up when Docker reports it as running, down when it has stopped or exited, and unknown when its host cannot be reached.
• Kubernetes: the container is up when Kubernetes reports it as ready, down when it has terminated, and otherwise unknown.
• Network or manual: DockDash probes the service’s check port. HTTP and HTTPS services accept any response below 500; if the HTTP request fails, DockDash falls back to a TCP connection. Other protocols use a TCP connection check.

HEALTH_CHECK_INTERVAL controls the polling interval. Status transitions can produce down and recovery notifications when Apprise is configured.

Health history

Each completed check is stored in SQLite and summarized as uptime history for the last 1, 7, or 30 days. HEALTH_HISTORY_TTL_DAYS controls retention, and DISABLE_HEALTH_HISTORY stops recording history when you do not need it.

Container resources

For Docker and Kubernetes services, DockDash samples CPU and memory usage. Docker also reports network and disk I/O. Historical CPU and memory samples make short spikes visible after the current reading has changed.

Kubernetes resource readings require the cluster metrics API. Memory percentage also requires a memory limit on the container; without a limit, DockDash can show usage but cannot calculate a meaningful percentage.

Resource history uses the same HEALTH_HISTORY_TTL_DAYS retention period as health history.

Threshold notifications

Optional CPU and memory thresholds send warning and recovery notifications. CPU must remain above its threshold for SPIKE_DURATION_THRESHOLD seconds; memory alerts as soon as it crosses its threshold.

See Notifications and the configuration reference for the related settings.