Use Amazon Managed Prometheus and Grafana for EKS cluster monitoring

[pbobov]

aws/arcgis-enterprise-k8s/enterprise-k8s-aws-organization workflow configures a monitoring subsystem that consists of:

• An SNS topic for monitoring alerts.
• A CloudWatch alarm that monitors the ingress ALB target groups and post to the SNS topic if the number of unhealthy instances in nonzero.
• A CloudWatch dashboard that displays the CloudWatch alerts, metrics, and container logs of the deployment.

The monitoring subsystem provides basic monitoring. However, a typical Kubernetes cluster monitoring solution involves the use of Prometheus and Grafana.

aws/arcgis-site-core/site-k8s-cluster-aws workflow should configure Amazon Managed Service for Prometheus and Amazon Managed Grafana for EKS cluster monitoring.

[pbobov]

Amazon Managed Prometheus and Grafana services are not supported in all the AWS regions (not supported in us-west-1 !).
Configuring networking and security for these services is complex, especially in disconnected VPC.
It looks like deploying Prometheus and Grafana in AKS clusters is a path of least resistance.