Deploy the prometheus-federation helm chart on Kubernetes

Question

Pulumi · Accepted Answer

To deploy the Prometheus Federation Helm chart on a Kubernetes cluster using Pulumi, we'll follow these steps:

Setup the Pulumi Project: Begin by creating a new Pulumi project if you don't already have one or use an existing Pulumi project.
Configure Kubernetes Provider: Ensure that Pulumi can communicate with your Kubernetes cluster. This typically involves setting up a kubeconfig file that Pulumi uses to interact with your cluster. Make sure you have kubectl configured properly and that it can access your cluster.
Write the Deployment Script: We'll write a TypeScript program that declares the deployment of the Prometheus Federation Helm chart using the Pulumi Kubernetes Provider.
Preview and Deploy: Before deploying the Helm chart, you can use Pulumi’s preview feature to see what changes will be applied to your infrastructure. Once you're ready, you'll run the pulumi up command to deploy the Helm chart to your cluster.

In our deployment script, we'll use the kubernetes.helm.v3.Chart resource from Pulumi’s Kubernetes provider, which is an abstraction for Helm charts. This allows us to deploy the Prometheus Federation Helm chart as if we were using Helm directly, but with all the benefits of Pulumi's infrastructure as code approach.

Here's a program that does just that:

import * as k8s from "@pulumi/kubernetes";

// Define the settings for the Prometheus Federation Helm chart.
// Replace `<version>` with the version number of the chart.
const prometheusFederationChart = new k8s.helm.v3.Chart("prometheus-federation", {
    chart: "prometheus-federation",
    version: "<version>", // Specify the chart version you want to deploy
    // Replace `<repo>` with the name of the repository where the chart is located, if needed.
    // If the chart is in the official Helm repo, it's not necessary to specify `repo`.
    fetchOpts:{
        repo: "<repo-url>", // Specify the Helm repository URL
    },
    values: {}, // If you have any configuration values, specify them here
});

// Export the base URL for the Prometheus dashboard, so we can easily access it
export const prometheusUrl = prometheusFederationChart.getResourceProperty(
    "v1/Service", "<namespace>", "prometheus-federation-server", "status"
).apply(status => `http://${status.loadBalancer.ingress[0].hostname}:${status.port}`);

Explanation:

The @pulumi/kubernetes package is imported to interact with Kubernetes resources.
We create a new Helm chart resource using new k8s.helm.v3.Chart, and we give it the name prometheus-federation.
Within this resource, we specify the chart and version properties, which Pulumi will use to locate the Prometheus Federation chart and ensure it deploys the correct version.
The fetchOpts.repo field is used if the Helm chart is hosted on a custom Helm repository; provide the URL where the chart can be found.
If the chart requires additional custom values, they would be specified in the values object. This allows you to configure the installation to suit your specific needs, all as part of your Pulumi program.
We then export the URL to the Prometheus dashboard. The getResourceProperty method is used to fetch information about the deployed service, specifically the hostname and port number. It assumes the service type is LoadBalancer and that it has an ingress with a hostname set. Modify this part based on your actual service type and expected output.
Replace <version>, <repo>, and <namespace> with appropriate values for your setup.

To actually deploy this to your Kubernetes cluster, run the following commands in your terminal:

pulumi up

This command will start the deployment process. Pulumi will show you a preview of the resources that will be created. You can confirm the action, and Pulumi will proceed with the deployment. After the deployment is complete, Pulumi will output the URL to access the Prometheus dashboard, which you can use to verify the success of the deployment.

Please replace placeholders such as <version>, <repo-url>, and <namespace> with actual values that pertain to your environment before running the deployment.