This is a basic demonstration of how it functions, based on the following tutorial: https://medium.com/@KimiHuang/monitor-your-eks-cluster-by-amazon-managed-service-for-prometheus-amp-f009ba149cab
aws amp create-workspace --region=eu-west-1 --alias amp_poc
Now, let's perform some ClickOps. Open your AMP workspace from the AWS console, copy the Helm configuration file, and paste it into the amp_ingest_override_values.yaml file.
Use eksctl to create/delete the EKS cluster.
Create:
eksctl create cluster -f eks.yaml
Delete:
eksctl delete cluster --name=eks-poc --region=eu-west-1
Execute the "zebra.sh" bash script.
chmod +x zebra.sh
./zebra.sh
Let’s use port-forward command to forward amp-server traffic to localhost:3000.
kubectl port-forward -n observability service/prometheus-for-amp-server 3000:80
Using browser to visit http://localhost:3000/
This script only creates the roles: amp-ebs-csi-iam-role and amp-iamproxy-ingest-role. If you run the script again, these roles won't be updated. If you plan to start from scratch, please delete these roles manually. Failing to do so may result in your Prometheus PVC being stuck in a pending state.
If you're adding a Prometheus data source to Amazon Managed Grafana, be sure to use Sigv4 for authentication and ensure that Grafana operates within the same VPC.