Skip to content

mcp.md

Latest commit

 

History

History
101 lines (68 loc) · 3.36 KB

mcp.md

File metadata and controls

101 lines (68 loc) · 3.36 KB
Raw

+++ title = 'MCP' weight = 2 cpu_total = '3 CPU cores' memory_total = '5 GiB' +++

The MCP overlay extends the core stack with the StreamsHub MCP server, giving AI assistants (Claude Code, VS Code Copilot, etc.) read-only access to the Strimzi-managed Kafka cluster.

Quick-Start Install

curl -sL https://raw.githubusercontent.com/streamshub/developer-quickstart/main/install.sh | OVERLAY=mcp bash

Manual Install

If you prefer step-by-step control, the MCP overlay uses overlays/mcp instead of overlays/core:

# Phase 1 — Operators and CRDs
kubectl apply --server-side --force-conflicts -k 'https://github.com/streamshub/developer-quickstart//overlays/mcp/base?ref=main'

# Optionally, wait for the operators to be ready
kubectl wait --for=condition=Available deployment/strimzi-cluster-operator -n strimzi --timeout=120s
kubectl wait --for=condition=Available deployment/apicurio-registry-operator -n apicurio-registry --timeout=120s
kubectl wait --for=condition=Available deployment/streamshub-console-operator -n streamshub-console --timeout=120s

# Phase 2 — Operands and MCP server
kubectl apply -k 'https://github.com/streamshub/developer-quickstart//overlays/mcp/stack?ref=main'

Uninstall

curl -sL https://raw.githubusercontent.com/streamshub/developer-quickstart/main/uninstall.sh | OVERLAY=mcp bash

What Gets Added

On top of the components provided by the core overlay, the MCP overlay adds:

Component Namespace Description
StreamsHub MCP server streamshub-mcp MCP server for AI-assisted Kafka cluster management
Sensitive Role kafka Grants access to TLS certificates and pod metrics

The MCP server provides read-only tools for inspecting and troubleshooting Strimzi-managed Kafka clusters via the Model Context Protocol.

Resource Requirements

The MCP overlay requires at least {{< param cpu_total >}} and {{< param memory_total >}} of allocatable cluster resources. This includes the resources for the core stack plus the MCP server listed above.

Accessing the MCP Server

Port-forward to the MCP server:

kubectl port-forward -n streamshub-mcp svc/streamshub-strimzi-mcp 8085:8080

The MCP endpoint is available at http://localhost:8085/mcp.

Verify the Installation

Confirm the MCP server is running:

kubectl get deployment -n streamshub-mcp streamshub-strimzi-mcp

The deployment should show 1/1 ready replicas. Check the health endpoint:

kubectl port-forward -n streamshub-mcp svc/streamshub-strimzi-mcp 8085:8080 &
curl -s http://localhost:8085/q/health/ready

Troubleshooting

MCP Server Not Starting

If the MCP server deployment is not becoming ready:

# Check pod status
kubectl get pods -n streamshub-mcp

# Check logs
kubectl logs -n streamshub-mcp deployment/streamshub-strimzi-mcp

# Verify Strimzi CRDs are installed
kubectl get crd kafkas.kafka.strimzi.io

Common causes:

  • Strimzi operator not installed — the MCP server requires Strimzi CRDs to exist
  • Insufficient cluster resources — check node resource availability
  • Image pull errors — verify the cluster can pull from quay.io