Monitoring Your SUSE Virtualization Environment

In today’s dynamic IT landscapes, gaining deep visibility into your hybrid cloud infrastructure is key to ensuring optimal performance and rapid troubleshooting of your virtual machines (VMs). If you’re looking for ways to modernize your infrastructure and reduce costs associated with solutions like VMware and Nutanix, this guide will walk you through how to set up comprehensive monitoring for your SUSE Virtualization environment, which offers containerized workloads in datacenters, public clouds, AI, and at the edge. From installing the SUSE Observability Agent to creating custom dashboards and scraping specialized KubeVirt metrics, you’ll gain insights into how to operate VMs in a cloud-native virtualization environment.

Monitoring your SUSE Virtualization cluster allows you to:

• Gain visibility: Track the performance of your virtual machines (VMs) and underlying Kubernetes infrastructure.
• Troubleshoot quickly: Identify issues before they impact production.
• Customize insights: Create tailored views and dashboards focusing on the metrics that matter most to your operations.

While you can monitor your cluster using the SUSE Observability Agent as you would with any standard Kubernetes setup, integrating additional KubeVirt metrics offers deeper insights into your virtualization environment.

Before you start, ensure you have the following in place:

• SUSE Observability Instance: A running instance to collect and display your metrics.
• SUSE Virtualization Cluster: Your virtualization environment must be up and running.
• Environment Configuration: Make sure your .env file is correctly configured (refer to the main README for setup details).

Confirm that you can connect to your cluster by running:

cd suse/virtualization/vmi-metrics
eval $(task common:shell-env)
kubectl get nodes

The first step involves setting up a Kubernetes StackPack instance. The instance name is determined by your CLUSTER_NAME environment variable. Simply run:

task install-stackpack-instance

Deploy the SUSE Observability Agent to start collecting metrics:

task common:deploy-observability-agent

After deployment, log in to your SUSE Observability instance. You should now see your cluster synced and begin receiving data.

SUSE Observability provides the flexibility to create custom views so you can focus on key components of your virtualization environment. This is particularly useful for isolating data on virtual machines and their controllers.

Using STQL queries, you can filter the data for different components:

• VirtControllers:

  (label IN ("kubevirt.io:virt-control") AND type = "pod")
  
• VirtualMachines:

  (label IN ("kubevirt.io:virt-launcher") AND type = "pod")
  

💡 Pro Tip: Star a view to pin it to the left menu for quick access.

For more details, check out the custom view documentation.

Your SUSE Virtualization environment exposes metrics through a Kubernetes service called kubevirt-prometheus-metrics. These metrics can be scraped and sent to SUSE Observability using the OpenTelemetry collector.

For a complete list of available metrics, refer to the KubeVirt metrics documentation.

Below is a sample configuration for the Prometheus receiver to scrape KubeVirt metrics:

config:
  receivers:
    prometheus:
      config:
        scrape_configs:
        - job_name: 'kubevirt-metrics'
          scrape_interval: 30s
          tls_config:
            insecure_skip_verify: true
          scheme: https
          kubernetes_sd_configs:
            - role: endpoints
              selectors:
                - role: endpoints
                  field: "metadata.name=kubevirt-prometheus-metrics"

Deploy the preconfigured SUSE Observability OpenTelemetry collector with:

task deploy-kubevirt-otel-collector

Once deployed, navigate to the Metrics page in SUSE Observability. All collected KubeVirt metrics begin with the prefix kubevirt_.

Visualizing your VM metrics is essential for effective monitoring. SUSE Observability allows you to create custom charts that can be adapted from pre-existing Grafana dashboards, such as the Harvester Grafana Dashboard and the KubeVirt Monitoring Dashboard.

Below are examples of charts you might consider, grouped by their function:

SUSE Observability uses STQL queries to bind metrics to specific topology components. The key is mapping metric tags to your component’s tags.

Given the Grafana PromQL query:

kubevirt_workqueue_unfinished_work_seconds{job=~".*kubevirt.*", instance=~"$instance"}

The equivalent query for a SUSE Observability chart would be:

kubevirt_workqueue_unfinished_work_seconds{k8s_cluster_name="${tags.cluster-name}", k8s_pod_name="${name}"}

Here’s an example of a complete chart configuration:

- id: -120
  name: Unfinished Work
  queries:
    - expression: kubevirt_workqueue_unfinished_work_seconds{k8s_cluster_name="${tags.cluster-name}", k8s_pod_name="${name}"}
      alias: '${name}'
  scope:  type in ("pod") and label = "kubevirt.io:virt-controller"
  identifier: urn:stackpack:kubevirt:shared:metric-binding:work-queue-unfinished-work
  unit: s
  chartType: line
  priority: high
  enabled: true
  layout:
    metricPerspective:
      tab: Virtualization
      section: Work Queue
    componentSummary:
      weight: 3
  _type: MetricBinding

Once you have defined your custom charts, apply them to your SUSE Observability instance with:

task upload-metric-bindings

Now, when you navigate to a virt-controller or virt-launcher pod’s metrics page, you will see the new Virtualization tab showcasing your custom dashboards.

By following these steps, you’ve set up an advanced monitoring environment for your SUSE Virtualization cluster. From deploying the SUSE Observability Agent and configuring custom views to scraping specialized KubeVirt metrics and creating tailored dashboards, you now have the tools to maintain peak performance and swiftly address issues within your virtual infrastructure.

Monitoring like this ensures that you capture all the critical details of your SUSE Virtualization environment, enabling you to keep a close eye on both your virtual machines and the underlying Kubernetes infrastructure. Happy monitoring! 🚀

The 258% ROI of SUSE Rancher Prime with Virtualization*

SUSE Rancher Prime with Virtualization delivers a compelling return on investment. Through IT cost savings, staff productivity gains, performance improvements and business enablement, IDC projects that interviewed organizations will achieve benefits worth an annual average of $3.4 million, which result in a three-year ROI of 258%.*

Download the full IDC paper to learn how SUSE Rancher Prime with Virtualization can transform your cloud native journey with business value.

You can find his github repo for this blog here

* IDC Business Value White Paper, sponsored by SUSE, The Business Value of SUSE Rancher Prime with Virtualization, February 2025 | IDC #US52704924