Chapter 2. Monitoring RHMAP with Cockpit
2.1. Overview
System resources of nodes and containers in the Core and MBaaS on OpenShift 3 can be monitored and managed using Cockpit.
Cockpit is a system administration tool, that provides insights into how nodes and containers are performing. It lets you monitor current values and adjust limits on system resources, control lifecycle of container instances, and manipulate container images. For more information about Cockpit, refer to the official web site of the Cockpit Project and its Documentation.
2.2. Installation
For most OpenShift 3 instances, Cockpit is most likely already installed on all nodes. This is not the case if your nodes use the RHEL Atomic Host, where Cockpit needs to be installed manually.
To check whether Cockpit is installed in your OpenShift cluster, try visiting the URL of the Cockpit web interface:
http://<master node host>:9090
If there’s no response to the request, Cockpit is most likely not installed.
2.2.1. Installing Cockpit Manually
Install Cockpit on nodes.
The following three steps must be repeated for each node you wish to monitor in your OpenShift cluster.
Log in to the node.
ssh <node host>
Install Cockpit packages.
yum install cockpit cockpit-docker
Enable and start the Cockpit service.
systemctl enable cockpit.socket systemctl start cockpit.socket
Create a Cockpit system user on master.
To log in to the Cockpit web interface, you will have to provide the username and password of an operating system user existing on the OpenShift master node. This guide refers to this user as the Cockpit system user. To allow Cockpit to access system resources, perform operations on Container and Kubernetes resources, the Cockpit system user must:
-
be in the
dockergroup; -
be able to log in to other nodes using
ssh; - be able to perform Kubernetes operations.
Create the Cockpit system user on the master node, or modify an existing user to have the necessary privileges.
-
be in the
2.3. Viewing the Containers on an Openshift Node
Navigate to the Cockpit dashboard for a node in a web browser (port 9090 by default) and log in as the Cockpit system user. To see all containers deployed on that node, click Containers in the left-hand side menu.
You can filter the list to only display running containers, using the dropdown menu above the list of containers. This view lets you see the RAM and CPU usage of all running containers.
If you select an MBaaS node, you will see the containers for all MBaaS components. Clicking on a container will show the current logs, CPU shares, and RAM usage. In the Tools menu on the left hand side, you can get terminal access into the node for further investigation.
2.4. Viewing Multiple Hosts Simultaneously
Cockpit can connect to multiple hosts from a single Cockpit session. This can be useful to compare resource usage of two or more machines in the same dashboard. See Multiple Machines in the Cockpit documentation for more information.

Where did the comment section go?
Red Hat's documentation publication system recently went through an upgrade to enable speedier, more mobile-friendly content. We decided to re-evaluate our commenting platform to ensure that it meets your expectations and serves as an optimal feedback mechanism. During this redesign, we invite your input on providing feedback on Red Hat documentation via the discussion platform.