ACM Agent

This is a multi-agent system for troubleshooting Kubernetes applications. Its initial implementation focuses on diagnosing the Open (or Advanced) Cluster Management environment, but it can be customized for other products as well.

Environment

We set up an Open Cluster Management environment to demonstrate how it works.

• Installation

Install clusteradm CLI tool Run the following command to download and install the latest clusteradm command-line tool:

curl -L https://raw.githubusercontent.com/open-cluster-management-io/clusteradm/main/install.sh | bash

Setup hub and managed cluster Run the following command to quickly setup a hub cluster and 2 managed clusters by kind.

curl -L https://raw.githubusercontent.com/open-cluster-management-io/OCM/main/solutions/setup-dev-environment/local-up.sh | bash

Organize Agents

Task 1: Interact with the Kubernetes Environment

• User: The user who ask questions and give tasks
• Executor: Execute the code written by the 'Engineer' and report the results back to them
• Engineer: Analyze the intent of the user or planner to write a sequence of shell commands or scripts

Demos

• Check the status of managedclusters

• Retrieve the resource usage of kafka cluster

  

Task 2: Add Knowledge Advisor for ACM

• Planner - Kubernetes planner, responsible for making a detailed plan to accomplish a specific task within a Kubernetes environment

• Advisor - The knowledge repository where you can find solutions and ideas for addressing any multi-cluster issues

Task3: Orchestrate all agents within the system

• Manager - orchestrates the workflow between agents

Demo: The Status of Cluster Unknown

• Scenario 1: Cluster1 - Make the bootstrap hub kubeconfig invalid

  # kubectl edit secret bootstrap-hub-kubeconfig -n open-cluster-management-agent --context kind-cluster1
  # kubectl edit secret hub-kubeconfig-secret  -n open-cluster-management-agent --context kind-cluster1
  
  kubectl delete secret bootstrap-hub-kubeconfig -n open-cluster-management-agent --context kind-cluster1
  kubectl delete secret hub-kubeconfig-secret  -n open-cluster-management-agent --context kind-cluster1

  kubectl get mcl cluster1 --context kind-hub
  python main.py "why the status of cluster1 is unknown?"

  

• Scenario 2: Disable the Klusterlet agent and the registration agent

  • Scale these 2 agents to 0

  kubectl scale deployment klusterlet -n open-cluster-management --replicas=0 --context kind-cluster2
  kubectl scale deployment klusterlet-registration-agent -n open-cluster-management-agent --replicas=0 --context kind-cluster2

  • Troubleshooting the unknown issue

  kubectl get mcl cluster2 --context kind-hub
  python main.py "why the status of cluster2 is unknown"