MCP for DevOps

GitHub Actions, PagerDuty, Splunk Integration

Build a unified DevOps platform where Claude orchestrates your entire incident response, deployment, and observability workflow. Connect GitHub Actions for CI/CD, PagerDuty for incident management, and Splunk for real-time log analysis—all through a single conversational interface.

Find DevOps Servers Back to Learn

Prerequisites

• GitHub repository with Actions enabled
• PagerDuty account with API access (Pro or higher)
• Splunk instance with indexing enabled
• CuratedMCP account and Claude Desktop installed

Building Your DevOps Stack

Set up GitHub MCP server

Start by configuring the GitHub MCP server to access repositories, workflows, and issues.

Generate a GitHub personal access token with 'repo' and 'workflow' scopes
Add it to your CuratedMCP configuration under the GitHub server
Test the connection by listing recent workflows

Configure PagerDuty MCP integration

Connect PagerDuty for incident management and on-call scheduling.

Create a PagerDuty API token in your account settings
Configure the PagerDuty MCP server with your token
Enable incident creation, escalation, and scheduling tools

Connect Splunk for observability

Integrate Splunk to query logs, metrics, and dashboards in real-time.

Set up a Splunk API token with search and data access permissions
Configure the Splunk MCP server with your instance URL and token
Create saved searches for common queries (errors, latency, deployment status)

Generate unified configuration

Use CuratedMCP to generate a single configuration for all three servers.

Visit /marketplace and add GitHub, PagerDuty, and Splunk MCP servers to your stack
Click 'Configure Stack' to generate a combined Claude Desktop config
Copy the JSON snippet into your Claude Desktop configuration file

Test end-to-end workflows

Verify that Claude can access all three services seamlessly.

Ask Claude: 'Show me failed GitHub Actions in the last 24 hours'
Ask: 'List active incidents from PagerDuty with pending escalations'
Ask: 'Search Splunk for errors in the API logs since 6am'

Deploy and automate

Use Claude to automate routine DevOps tasks across all three platforms.

Claude can trigger GitHub Actions, create PagerDuty incidents, and query logs—all in one conversation
Save scripts and workflows that Claude creates for you
Use CuratedMCP to share this stack with your team

Real-World Use Cases

Incident Response Automation

When a Splunk alert triggers, Claude queries logs, creates a PagerDuty incident, and kicks off a remediation GitHub Action—all in seconds.

Deployment Coordination

Plan a deployment: Claude checks recent PagerDuty schedules (who's on-call), verifies GitHub CI is passing, and coordinates with the team.

Post-Incident Reviews

After an incident, Claude pulls Splunk logs, PagerDuty timeline, and related GitHub commits to generate a comprehensive incident report.

Proactive Monitoring

Claude monitors Splunk dashboards and Splunk alerts for emerging issues. If threshold exceeded, it can immediately notify the on-call engineer via PagerDuty.

Configuration Snippet

After configuring all three servers on CuratedMCP, your Claude Desktop config will include sections like this:

{
  "mcpServers": {
    "github": {
      "command": "npx",
      "args": ["github-mcp-server"],
      "env": {
        "GITHUB_TOKEN": "ghp_your_token_here"
      }
    },
    "pagerduty": {
      "command": "npx",
      "args": ["pagerduty-mcp-server"],
      "env": {
        "PAGERDUTY_API_TOKEN": "u+your_token"
      }
    },
    "splunk": {
      "command": "npx",
      "args": ["splunk-mcp-server"],
      "env": {
        "SPLUNK_HOST": "https://your-splunk.com",
        "SPLUNK_TOKEN": "your_api_token"
      }
    }
  }
}

Best Practices

Use Environment Variables

Never hardcode API tokens. Store them in environment variables or your system keychain, then reference them in your MCP configuration.

Version Your Configuration

Store your MCP config files in version control (exclude secrets). This makes it easy to sync configurations across your team and roll back if needed.

Test in a Staging Environment

Before enabling destructive actions (incident creation, deployments) in production, test the integration with staging services.

Next Steps

Browse the CuratedMCP marketplace for GitHub, PagerDuty, and Splunk MCP servers
Generate a unified stack configuration using CuratedMCP's Config Generator
Test individual integrations (query GitHub, list incidents, search Splunk)
Build a sample automation workflow in Claude to combine all three services
Share your stack with your team via CuratedMCP

Ready to Build Your DevOps Stack?

Explore DevOps MCP Servers

← Back to Learning Hub

Questions? See our setup guide.