How can Amazon Managed Service for Prometheus automatically collect metrics from Amazon EKS in TypeScript

Question

Pulumi · Accepted Answer

To set up automatic metric collection from Amazon EKS using Amazon Managed Service for Prometheus (AMP), you would create a workspace in AMP, configure a service discovery to find EKS cluster metrics endpoints, and deploy Prometheus server agents that scrape the metrics, sending them to the AMP workspace.

Here is an outline of the steps we will follow in the Pulumi program:

1. Create an AMP workspace which is a logical place where Prometheus can store metric data.
2. Configure your EKS cluster to expose the metrics endpoints if not already exposed.
3. Deploy a Prometheus server into your EKS cluster which will scrape the metrics from the cluster and send them to your AMP workspace.

The code below is a Pulumi program written in TypeScript, which will set up a workspace in AMP. It does not include the actual deployment of the Prometheus server on your EKS cluster, but lays out the necessary infrastructure:

```typescript
import * as pulumi from "@pulumi/pulumi";
import * as aws from "@pulumi/aws";

// Create an Amazon Managed Service for Prometheus (AMP) workspace.
const workspace = new aws.amp.Workspace("myWorkspace", {});

// Output the Prometheus workspace id.
export const workspaceId = workspace.id;
```

Before running this Pulumi program, you will need to set up your AWS and Pulumi CLI credentials. With this infrastructure code, you will have an AMP workspace ready for Prometheus metrics collection. After running this Pulumi code, you will still need to deploy Prometheus to your EKS cluster and configure it to send metrics to the created workspace.

To deploy Prometheus on your EKS cluster, you may consider using the Prometheus community Helm chart. This would involve setting up the Helm provider in Pulumi and crafting a `Chart` resource that points to the Prometheus Helm chart while configuring it to use your AMP workspace as the backend.

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

// Firstly, you need to set up an Amazon Managed Service for Prometheus workspace.
const workspace = new aws.amp.Workspace("myWorkspace", {});

// Deploy Prometheus to an existing Kubernetes cluster.
const cluster = new k8s.Provider("myeks", {
  kubeconfig: yourExistingEksClusterKubeconfig,
});

const chart = new k8s.helm.v3.Chart("prometheus", {
  namespace: "monitoring",
  chart: "prometheus",
  fetchOpts: { repo: "https://prometheus-community.github.io/helm-charts"},
  version: "14.11.1",
  values: {
    server: {
      remoteWrite: [
        { url: `https://${workspace.prometheusEndpoint}` }
        // Additional configuration for authentication, if required.
      ],
      // Additional Prometheus server configurations.
    },
    // Optionally, include other components like alertmanager, nodeExporter, etc.
  },
}, { provider: cluster });

// Export necessary information, such as the workspace id and Prometheus server URL.
export const workspaceId = workspace.id;
export const prometheusServerUrl = pulumi.interpolate`http://${chart.getResourceProperty("v1/Service", "monitoring/prometheus-server", "status").apply(s => s.loadBalancer.ingress[0].hostname)}`;
```

Remember, the above code should be included in your Pulumi program to successfully deploy a Prometheus server in your EKS cluster that automatically collects metrics and sends them to your AMP workspace. The exact values might need adjustments and additional parameters based on your EKS environment and monitoring requirements.

The program requires your EKS cluster `kubeconfig` to be provided where it's indicated in the code. Be sure to review the security and authentication requirements for remote write access to AMP. You may need to manage AWS credentials for this access. You can find more details about this setup in the [AWS AMP Documentation](https://docs.aws.amazon.com/prometheus/latest/userguide/AMP-onboard-create-workspace.html).