About the Monitoring Module

Cluster Overview Page

The The "Overview" page icon is a square divided into several other squares. Overview page of your cluster provides information about the resource consumption of all your platforms.

By default, the Cluster Overview page opens when you click the The "Monitoring" module icon is a heart with an electrocardiogram in it. Monitoring module.

The page gives you an overview of your cluster’s node consumption and reservation. It is composed of a two-column table describing the consumption of CPU CPU and RAM RAM resources.

This is the entry point for monitoring the overall capacity of your cluster and the health status of its nodes. Based on these information, you can adjust the limits and requests accordingly for each platform.

To have proper access to monitoring details, your administrator must have configured Saagie to isolate the workload between platforms. This is done by specializing your nodes per platform. If workload isolation has not been configured, the The "Monitoring" module icon is a heart with an electrocardiogram in it. Monitoring module will not be fully operational, as all your cluster nodes will be displayed on each platform without disassociation. For more information, see Node Isolation.
Example 1. Saagie installed in isolated and non isolated mode.

Here is the example of Saagie installed with two platforms. One installation with your isolated workload and the other without.

  • Saagie installed in isolated mode

  • Saagie installed in non-isolated mode

If you have four nodes in your cluster, the The "Overview" page icon is a square divided into several other squares. Cluster Overview page will show your platforms with their dedicated node(s) only. If a node is not labeled, it will not be considered for Saagie runs and will not appear.

monitoring overview cluster iso
Figure 1. The Cluster Overview page displays the platforms with only the node(s) dedicated to them.

If you have four nodes in your cluster, the The "Overview" page icon is a square divided into several other squares. Cluster Overview page will show all the cluster nodes on each of your platforms without dissociation, even if no resources are used on those nodes by your platform.

monitoring overview cluster niso
Figure 2. The Cluster Overview page displays all the cluster nodes on each platform without dissociation.

App, Job, and Pipeline Resource Consumption Graphs

The The "Overview" page icon is a square divided into several other squares. Overview and The "Instances" page icon is three overlapping squares. Instances pages for apps, jobs, and pipelines also include graphs that allow you to track consumption as the job, app, or pipeline runs on the node. It can help you quickly identify bottlenecks, debug jobs and apps going OOM Out Of Memory, and better optimize resource usage on your platform.

You can click the graph line to display tooltips with more information, and you can zoom in on a specific period.

monitoring graph consumption app

Resource Details Page

The open external Resource Details page is an external link that will take you to the Saagie Resources Monitoring (SRM) page. SRM gives you an overview of your cluster’s resource consumption through a series of graphs.

SRM is based on Grafana. It is a set of graphs providing a global view of RAM and CPU consumption for nodes, jobs, and apps of your cluster.

For more information, see About Saagie Resources Monitoring.