System Overview

Real-time monitoring of the host and Docker infrastructure from the System page — not only per-container stats.

Metrics

Metric	Details
CPU	Load averages (1 / 5 / 15), usage per core
Memory (RAM)	Used, free, cache, and swap
Disk	Capacity and read/write activity
Network	Upload and download rates
GPU	Shown when available on the host
Processes	Top processes on the machine

Use this page to spot disk pressure, memory exhaustion, and runaway processes before deploys fail.

Purge

Purge is a Docklift-scoped maintenance action that requires your account password (step-up auth). JWT alone is not enough.

What it does:

• Removes unused docklift-* app image tags that are outside each project’s keep-2 set (current + previous successful deploy), skipping tags still used by running containers
• Clears all BuildKit cache (docker builder prune -af) so disk can recover after many redeploys
• Returns 409 if any deployment is still in progress (so in-flight build tags are not deleted)

What it never does:

• Host-wide docker system prune
• Delete non-docklift-* images (Postgres, nginx, etc.) or foreign containers
• Wipe OS caches, journals, apt caches, or /tmp

After every successful deploy, Docklift also keeps only two images per project service and prunes unused BuildKit cache automatically. See Deployment.

Control plane actions

Action	Effect
Reset	Restart Docklift platform services
Reboot	Restart the server

Dangerous host actions from the terminal strip also require step-up password confirmation. Cancel aborts — the action does not continue.

Logs

Dashboard → Logs streams platform containers over SSE:

• Backend
• Frontend
• Nginx (dashboard gateway on :8080)
• Nginx Proxy (public :80/:443)

From the host:

docker logs docklift-backend -f --tail 100
docker logs docklift-nginx -f --tail 100
docker logs docklift-nginx-proxy -f --tail 100

Health

curl -s http://SERVER_IP:8080/api/health

On the server itself you can also use http://127.0.0.1:8080/api/health.