Container Monitoring

Explore → Containers gives you one view of every Docker container running on the hosts where you’ve installed a Log Forwarder with container monitoring turned on.

Turning it on

Container monitoring isn’t automatic — it comes from two Log Forwarder modules you turn on for a group of forwarders:

• Container Stats — CPU, memory, and network usage per container.
• Docker Logs — each container’s logs, tailed via the host’s system log.

See Forwarder Management for how to turn these on and what each one needs on the host. Once enabled, containers on that host appear here with no further setup.

Host View and Container View

The page has two tabs:

• Host View — one row per host, with a rollup of how many containers it’s running and whether it’s currently online.
• Container View — one row per container across every host, so you can search and scan the whole fleet at once.

Both are searchable by container name, image, or host name. Click any row to open its full detail.

What you see per container

• Status:
  • Online — the container is reporting normally.
  • Stale — GridNMS knows the container but hasn’t had a recent sample from it.
  • Forwarder offline — the forwarder on the container’s host isn’t reporting, so the container’s own state is unknown.
  • Retired — the container stopped reporting about an hour ago or more.
• Image and the host it’s running on.
• Trend sparklines for CPU, memory, and network — so you can spot a container that’s climbing without opening it.

Opening a container shows its full CPU, memory, and network charts plus its recent logs, all in one drawer.

Retired containers

Containers come and go — a rebuilt service or a finished job leaves its old container behind. So the list stays useful, a container that has stopped reporting for about an hour is marked Retired and hidden by default.

Turn on Show retired at the top of the page to include them — handy when you’re looking for something that used to run on a host. Retired containers are removed for good after 14 days. If one starts reporting again before then, it comes straight back as a normal container.

Alerting on a container

Open a container and use its Thresholds panel to get alerted when it runs hot. Click Add, pick the Metric (CPU, memory, or network/disk I/O), set the Condition and Value, and choose a Severity. You can switch a threshold between Active and Inactive without deleting it, and edit or remove it from the same panel.

For alerting on what a container logs rather than what it uses, write a Detection over its logs.

Next steps

• Forwarder Management — turn on Container Stats and Docker Logs.
• Installing a Log Forwarder — get a forwarder running on your Docker hosts.