Monitoring and cleanup
CPU, memory and disk charts, server statuses, disk alerts and Docker cleanup.
Warden checks every server once a minute over one SSH connection: CPU, memory, disk, and Docker’s containers and disk use.
Statuses
Section titled “Statuses”| Status | Meaning |
|---|---|
| Connected | Reachable, with containers running |
| Idle | Reachable, with nothing running (or no Docker) |
| Offline | Three checks in a row failed |
When a server that was reachable goes Offline, and when it comes back, Warden can notify you.
Resources
Section titled “Resources”The server page’s Resources tab shows the last 24 hours: CPU (with load averages), memory and disk, as 5-minute averages with peak and average. Memory counts what’s really in use (total minus what Linux can free). The Docker disk chart, checked every 10 minutes, splits Docker’s space into images, volumes, build cache and containers, and shows how much is reclaimable.
Readings are kept for 7 days.
Containers lists what’s running: state, CPU, memory and status.
Disk alerts
Section titled “Disk alerts”When the disk passes 90% full, Warden sends a Disk almost full alert once. It alerts again only after usage has dropped below 85% and climbed back.
Run cleanup
Section titled “Run cleanup”Run cleanup (DevOps and Admin) frees disk space Docker no longer uses. It shows what it would free first, then removes, in order:
- Stopped containers.
- Dangling images (untagged leftovers).
- Build cache.
- Unused networks.
It never removes volumes (your databases’ data) or tagged images (which rollbacks use). You can close the window while it runs.