MCP Packs
DevOps/SRE Pack
Operate services with cloud, cluster, and monitoring MCP servers.
DevOps/SRE Pack
Operate your infrastructure and respond to incidents faster by connecting cloud platforms, container orchestration, and observability tools through MCP.
Recommended Servers
| Server | Purpose |
|---|---|
| Kubernetes MCP | Query pods, logs, events, deployments |
| AWS MCP | Describe resources, invoke runbooks |
| Prometheus MCP | Query metrics and alerts |
| Terraform MCP | Plan and apply infrastructure changes |
| GitHub MCP | Manage repos, PRs, and deployments |
Example Configuration
{
"mcpServers": {
"kubernetes": {
"command": "npx",
"args": ["-y", "@kubernetes/mcp-server"],
"env": { "KUBECONFIG": "${KUBECONFIG}" }
},
"aws": {
"command": "npx",
"args": ["-y", "@aws/mcp-server"],
"env": {
"AWS_ACCESS_KEY_ID": "${AWS_ACCESS_KEY_ID}",
"AWS_SECRET_ACCESS_KEY": "${AWS_SECRET_ACCESS_KEY}"
}
},
"prometheus": {
"command": "npx",
"args": ["-y", "@prometheus/mcp-server"],
"env": { "PROMETHEUS_URL": "${PROMETHEUS_URL}" }
}
}
}Common Workflows
- Incident triage: Fetch logs from Kubernetes, check Prometheus alerts, then roll back via AWS or kubectl
- Capacity planning: Query resource utilization, analyze trends, plan scaling
- Deployment verification: Check pod status, review metrics, validate health endpoints
Safety First
Use read-only credentials where possible. Restrict write operations to specific namespaces or resources. Never expose production secrets in MCP configs.