SR-IOV Network Monitoring Helm Chart

Basic Introduction

This is a Helm chart package to manage SR-IOV Network Monitoring deployment on a Kubernetes cluster.

Prerequisites

• Ensure the Prometheus and Grafana Operators are set up in order to collect and visualise the exported logs.

Installation

Using NGC CLI and NGC Web UI

Please see the NGC Helm Charts documentation for general guidance.

Using Helm Dashboard

This Helm chart can be installed on your Kubernetes cluster using the Helm Dashboard UI once the Helm Dashboard Helm Chart is installed.

Using Helm

1. Login to the NGC registry

   helm registry login nvcr.io
   

2. Add repo on your local system

   helm repo add <your-repo-name> https://helm.ngc.nvidia.com/nvidia/holoscan-for-media/ '--username=$oauthtoken' --password=<API-KEY>
   

   where:

   • <your-repo-name> is a short name like holoscan-for-media to be used on your local system to indicate the chart repository on NGC specified by the URL
   • <API-KEY> is your NGC API Key

3. Update the repo or fetch particular chart

   helm repo update
   

   or

   helm pull <your-repo-name>/sriov-network-monitoring
   

   where:

   • <your-repo-name> is the short name you used in the helm repo add command above

4. Now you can see all the charts available

   helm search repo
   

5. You are ready to install SR-IOV Network Monitoring using its Helm chart

   Install using below command

   helm install --set metricsExporter.cgroup=<v1-or-v2> <release-name> <your-repo-name>/sriov-network-monitoring
   

   where:

   • <v1-or-v2> is the cgroup version of the cluster, which can be checked by running the command:
     cgroup=$(stat -fc %T /sys/fs/cgroup/) && cgroup=${cgroup/cgroup2fs/v2} && cgroup=${cgroup/tmpfs/v1} && echo $cgroup
     

   • <release-name> is a short name, like sriov-network-monitoring, which should be unique on your local system, to indicate this particular installation of this application
   • <your-repo-name> is the short name you used in the helm repo add command above

6. After installation you can check the status

   kubectl get all -o wide
   

7. Once the pod is running, the SR-IOV Network Dashboard become visible on Grafana UI

   Note: SR-IOV Network Dashboard will take 1-2 minutes to start showing the metrics

8. Other commands to manage the installed application

   helm status <release-name>
   helm get all <release-name>
   helm uninstall <release-name>
   

   where:

   • <release-name> is the short name you used in the helm install command above