Docker Monitoring

Our Docker LogicModule monitors resource usage within Docker containers. To do so it requires the installation of the “cAdvisor” monitoring container within your docker service.

Installing cAdvisor

To install cAdvisor, execute the following commands on your Docker host:

sudo docker run \
  --volume=/:/rootfs:ro \
  --volume=/var/run:/var/run:ro \
  --volume=/sys:/sys:ro \
  --volume=/var/lib/docker/:/var/lib/docker:ro \
  --publish=8080:8080 \
  --detach=true \
  --name=cadvisor \
  gcr.io/cadvisor/cadvisor:latest

Once cAdvisor is running, LogicMonitor will discover containers as instances and begin gathering data.

For more information, see https://github.com/google/cadvisor/

Running cAdvisor on RHEL and Centos

On RHEL and Centos systems you’ll need to add these install options to ensure proper operation of cAdvisor:

--privileged=true \
--volume=/cgroup:/cgroup:ro \

Running cAdvisor on Docker

On older versions of Docker (< 1.2) you may need to specify the root volume access option as:

--volume=//:/rootfs:ro \

Running cAdvisor on Docker Swarm

If you are deploying cAdvisor as a service in Docker Swarm (typically in global mode, one task per node), you must publish its port in host mode to bypass Swarm’s ingress routing mesh. Without this configuration, requests can be load balanced across nodes, resulting in inconsistent data collection. Each cAdvisor endpoint is node-local and reports only on containers running on the node where that task is hosted.

The following example shows how to configure host-mode publishing in docker-compose.yml:

ports:
  - target: 8080
    published: 8080
    protocol: tcp
    mode: host

This configuration ensures that the LogicMonitor Collector connects directly to the monitored node so the data returned accurately reflects the containers running on that specific node. For more information, see Bypass the routing mesh from Docker.