Deploy the prometheus-ci-metadata-exporter helm chart on Opensshift

Question

Pulumi · Accepted Answer

To deploy a Helm chart on OpenShift using Pulumi, you'll want to make use of the Pulumi Kubernetes package. This package allows you to manage Kubernetes resources, including deploying Helm charts. Below we will write a Pulumi TypeScript program to accomplish the goal of deploying the Prometheus CI Metadata Exporter Helm chart on an OpenShift cluster.

Here's a step-by-step explanation of the process:

1. First, you will need to have access to an OpenShift cluster and have your `kubectl` configured to connect to it.
2. We'll use the `kubernetes.helm.v3.Chart` resource to deploy the Prometheus CI Metadata Exporter Helm chart to your OpenShift cluster. This resource is an abstraction that represents a Helm chart, residing within the Kubernetes provider.
3. You will need to specify the necessary properties like chart name, repository, version, etc., to correctly identify and install the Helm chart you wish to deploy.
4. We'll set up the program structure with necessary imports and create an instance of `Chart`.
5. The namespace where you want to deploy the Helm chart will also be specified. If you don't have a specific namespace, the chart will be installed in the default namespace.
6. Values can be overridden or added for your particular Helm chart deployment by providing an object to the `values` property. These are equivalent to values you would usually set in a `values.yaml` file or via the command line with Helm.

Here's the program that follows these steps:

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

// Replace with your specific Helm chart details
const chartName = "prometheus-ci-metadata-exporter";
const chartVersion = "1.0.0"; // example version - use the actual chart version here
const chartRepositoryUrl = "https://my-helm-chart-repository.com/"; // replace with the actual chart repo URL

// Assuming we are installing the chart in the "monitoring" namespace.
// Make sure this namespace exists in your OpenShift cluster or change it to the desired one.
const namespace = "monitoring";

// Create a Helm Chart Resource using Pulumi. This will install the chart to your OpenShift cluster.
const prometheusChart = new k8s.helm.v3.Chart(chartName, {
    chart: chartName,
    version: chartVersion,
    fetchOpts: {
        repo: chartRepositoryUrl,
    },
    namespace: namespace,
    // Set any custom values for the Helm chart here.
    // For example:
    // values: {
    //   service: {
    //     type: "ClusterIP",
    //   },
    // },
});

// Export the URL for the Prometheus CI Metadata Exporter, if it creates an endpoint accessible outside the cluster.
export const prometheusServiceUrl = prometheusChart.getResourceProperty("v1/Service", "prometheus-ci-metadata-exporter-service", "status").apply(status => status.loadBalancer.ingress[0].hostname);

```

Explanation of what this program does:

- We import the Pulumi Kubernetes package, which allows us to interact with Kubernetes cluster resources.
- We define the Helm chart's configuration variables such as its name, version, and the repository's URL.
- We specify the namespace where the chart will be deployed. In this example, we're using a namespace called "monitoring".
- We create a new Helm chart resource. This resource references the specific chart we want to deploy, along with the version and other relevant options.
- Optionally, if the chart exposes a service with a LoadBalancer, we export the URL so that you know how to access the Prometheus exporter once it’s deployed. (Please note that the actual property paths may vary based on the chart, and you might need to adjust them accordingly).

Please review the chart details such as version and repository URL, and ensure that the namespace you're deploying to matches your setup. Once the program is in place, you would run `pulumi up` to initiate the deployment process. Pulumi will communicate with your OpenShift cluster and perform the necessary actions to deploy the Helm chart described by the code.