Docker monitoring

Containers watched,restarts and all.

Checkmate watches the containers your stack runs on: status, health checks, restart events and resource usage, next to the uptime monitors for the services inside them.

postgres-16Running · healthy
checked every 30 s
CPU
4.2%
Memory
312 MB
Uptime
18 days
0 restarts
Recent checks
09:41:02healthy
09:40:32healthy
09:40:02healthy
Container status

Up, healthy or quietly gone.

A container can exit at 3am and be restarted by its policy before anyone notices, or stay down until a customer does. Docker monitors track container state plus the health checks your images define, so either way there is a record and an alert.

If an image ships a HEALTHCHECK, Checkmate reads it. Unhealthy is the warning you get before stopped.

  • Container status tracking
  • Docker health check results
  • Alerts on the 12 channels you already use
Containershost: web-01
webUp · healthy
workerUp
redisUp · healthy
cronRestarting
legacy-apiExited
Restart events

Crash loops, caught in the act.

Restart events are reported alongside host metrics, so a container bouncing every few minutes stops hiding behind its restart policy.

The pattern is visible in history: one restart after a deploy is noise, 40 an hour is a bug.

  • Restart events in the timeline
  • History that separates blips from loops
Restart events3 in 15 min
09:38cron restarted (exit 137)
09:31cron restarted (exit 137)
09:24cron restarted (exit 137)
07:02web deployed · image updated
Pattern detected: restart loop on cron
Resources

Which container is eating the box.

Per-container resource usage, reported through the Capture agent next to the host's own CPU, memory and disk. When a machine runs hot, the answer to 'which one is it' is on the same page.

That beats SSHing in to run docker stats while the pager keeps buzzing.

  • Per-container resource usage
  • Host metrics on the same dashboard
  • Reported by the open-source Capture agent
Per-container usagehost CPU 64%
web
1.4 GB
postgres-16
2.1 GB
worker
820 MB
redis
190 MB
Who runs this

For everyone whose stack is a compose file.

Self-hosters

Your services live in containers on one or two boxes. Watch all of them without adding a heavyweight orchestration layer.

Small product teams

A VPS running a compose stack is a common production setup. Container health, host health and endpoint uptime in one dashboard covers it.

Homelab tinkerers

20 containers accumulate fast. Get a straight answer to which ones are up, unhealthy or stuck in a restart loop.

Scope

What it covers, and what it doesn't.

Covered

  • Container status and Docker health checks
  • Restart events
  • Per-container resource usage via Capture
  • Host-level metrics on the same dashboard
  • Alerts on 12 notification channels

Out of scope

  • Kubernetes control planes and pods: Checkmate watches containers, not orchestrators
  • Image registry and vulnerability scanning
  • Log collection from your containers
Under the hood

For the technically curious.

A first-class monitor type

Docker monitors sit next to HTTP, ping and the rest. Add one per container you care about, set the interval and route its alerts like any other monitor.

Health checks respected

The HEALTHCHECK your image already defines is the best signal of container wellbeing. Checkmate surfaces it instead of inventing a parallel definition of healthy.

Pairs with service checks

The container being up and the service inside answering are different facts. Run a Docker monitor next to an HTTP monitor for the same app, and know which layer broke.

Checkmate runs in Docker too

The whole platform ships as a small set of containers with a compose file. Monitoring containers from containers is the normal setup here.

FAQ

Frequently askedquestions.

Add a Docker monitor for a container. Checkmate tracks its status and health check results on the schedule you set. Restart events and resource usage appear alongside. Alerts route through your notification channels.

Yes. Restart events are recorded with the host metrics, so a restart loop shows up as a visible pattern and a status change fires your alert channels.

Yes, through the Capture agent, which reports per-container usage next to host-level CPU, memory and disk. Finding the container that is eating a machine takes one look.

No. Checkmate watches containers and hosts, not orchestrator control planes. If you run plain Docker or compose stacks, it fits. For Kubernetes-native observability you'll want a Kubernetes-native tool.

Yes. It ships as a small set of containers: pull the compose file, run it and the dashboard is up. About 1 GB of RAM on any Docker host is enough to start.

Yes. Open source under AGPL-3.0, self-hosted, with no container or host limits.

Get started

Every feature,no paywall.

Checkmate is open source under AGPL-3.0. Self-host it and this feature ships free, on your servers, with your data.