Monitoring CRI-O

eG Enterprise offers a specialized CRIO monitoring model that monitors each of the key indicators of the performance of Container Engine - such as the images, pods, containers etc.- and proactively alerts administrators to potential performance bottlenecks, so that administrators can resolve the issues well before end-users complain.

Figure 1 : The layer model of CRIO Container Engine

Each layer of Figure 1 above is mapped to a variety of tests, each of which report a wealth of useful information related to the docker server. Using these metrics, administrators can find quick and accurate answers to the following performance queries:

  • Is the CRIO service installed?
  • What is the current status of the CRIO service?
  • Are there any containers that were recently added to/removed from the CRIO Container Engine? Id so, which are those containers?
  • Are there any long running containers in the CRIO Container Engine? Which are the contianers that were running for a longer duration?
  • Is each container utilizing adequate memory, CPU and disk resources? Which container is utilizing the physical resources excessively?
  • How many images are available in the CRIO Container Engine?
  • Are there images that are not mapped to the containers? If so, which are those images that are not mapped to the containers?
  • Are there images that are not mapped to containers consuming too much of disk space? If so, which are those images?
  • How many Pods are currently available in the CRIO Container Engine?
  • How many Pods are currently running and how many are currently stopped?
  • Is each Pod utilizing adequate memory, CPU and disk resources? Which Pod is utilizing the physical resources excessively?
  • What is the uptime of each container?
  • Is the container available over the network?

The top three layers of the layer model are briefly discussed in the following sections;

The Operating System, Network and TCP layers of the CRIO monitoring model are similar to that of a Linux server and had been dealt in extensively in Unix and Windows Servers monitoring model.