Advanced Cluster Management Observability on ROSA

Deploying RHACM Observability on ROSA

This document provides a complete guide to deploying ACM Observability on a ROSA cluster with Thanos-based metrics storage using AWS S3

• Thanos-based metrics storage using AWS S3 for persistent, long-term metrics retention
• Grafana dashboards for multi-cluster observability and visualization
• Persistent metrics retention across managed clusters

Prerequisites

Required:

• An existing ROSA cluster (Classic or HCP)
• Advanced Cluster Management (ACM) 2.10+ installed and operational
• AWS CLI configured with appropriate permissions
• oc CLI authenticated to your ROSA cluster

Verify Prerequisites

Before proceeding, validate your environment meets all requirements:

If ACM is not installed: You must install ACM first. Follow the ACM Installation Guide before proceeding.

Architecture Overview

ACM Observability uses Thanos to collect, store, and query metrics from all managed clusters:

• Metrics are collected from managed clusters via the observability addon
• Thanos components (Query, Store, Compact) process and store metrics
• S3 provides durable, long-term storage for historical metrics
• Grafana provides visualization and dashboards

Step 1: Environment Setup

Set environment variables for the deployment:

Validate Environment Variables

Verify all required variables are set:

Step 2: Prepare AWS Infrastructure

Create S3 Bucket

Verify:

Create IAM Policy for S3 Access

Apply the Policy:

Verify:

Create IAM User

Verify:

Attach Policy to User

Verify:

Generate Access Keys

Step 3: Configure ACM Observability on OpenShift

Create Observability Namespace

Create Pull Secret

The observability components need a pull secret to download container images:

Verify:

Create Thanos Object Storage Secret

This secret contains S3 credentials for Thanos to store metrics:

Verify:

Validate secret content:

Step 4: Deploy MultiClusterObservability

Create the MultiClusterObservability custom resource to enable observability:

Expected Output: Resource created (status may show “Pending” initially)

Step 5: Verify Deployment

Monitor Deployment Progress

After the immediate creation of MultiClusterObservability, pods will start within 2 to 5 minutes, reach a running state within 5 to 10 minutes, and display the first metrics in 10 to 15 minutes.

Check All Pods Are Running

Expected Output:

Verify S3 Bucket Has Data

After 10-15 minutes, metrics should start appearing in S3:

Get Grafana Route

Test Grafana Access

Step 6: Access ACM Observability Dashboards

1. Via ACM Console:

   • Log into the OpenShift console
   • Navigate to ACM > Overview
   • Click on Grafana link in the Observability section

2. Direct Access:

   • Visit the Grafana URL from the previous step
   • Log in with your OpenShift credentials

3. Available Dashboards:

   • ACM - Clusters Overview: Health and status of all managed clusters
   • ACM - Resource Optimization: Resource usage and capacity planning
   • ACM - Applications: Application deployment metrics

Cleanup / Uninstall

To remove ACM Observability:

Additional Resources

• ACM Installation Guide: Installing ACM on OpenShift
• ACM Observability Documentation: Red Hat ACM 2.11 Observability
• Thanos Documentation: Thanos S3 Configuration
• ROSA Documentation: Red Hat OpenShift Service on AWS