Dynatrace MCP Server: Let AI Query Metrics, Problems, and Service Health

Why Give AI Access to Dynatrace?

Dynatrace monitors everything: application performance, infrastructure health, user experience, and service dependencies. When something goes wrong, you navigate between problems, service flow, logs, and metrics dashboards. During incidents, that context-switching costs minutes.

An MCP server for Dynatrace lets your AI agent query all of it directly. Check active problems, pull service metrics, investigate root causes, and review SLO status without opening the Dynatrace UI.

What Becomes Possible

Once your AI agent has Dynatrace access:

• "Are there any active problems in production right now?"
• "What's the response time trend for the checkout service?"
• "Show me the error rate for the payment API over the last hour"
• "Which services are affected by the current problem?"
• "What's the CPU usage on the database hosts?"
• "Are all SLOs green for this sprint?"

How It Works

1. Go to DataFaucet and enter your Dynatrace URL (e.g. https://your-env.live.dynatrace.com)
2. Log into your Dynatrace environment
3. Browse problems, services, metrics as you normally would
4. Each action becomes an MCP tool your AI can call
5. Deploy and connect to Claude, Cursor, or Windsurf

Works with Dynatrace SaaS and Managed. The AI agent gets whatever access your session provides.

Security Best Practices

• API tokens with minimal scope — use tokens scoped to read-only metrics and problems
• Environment-scoped — limit to specific management zones
• No configuration access — prevent the agent from modifying settings or rules
• IP restrictions — restrict token usage to known ranges
• Audit logs — Dynatrace records all API access

Use Cases by Role

SRE/On-call engineers: During incidents, ask "what problems are active right now?" and "which services are impacted?" without navigating the Dynatrace problem feed.

Backend developers: Performance checks mid-coding. "What's the P95 latency for the API I just changed?" without switching to a dashboard.

Platform engineers: SLO monitoring. "Are all production SLOs meeting target?" during sprint reviews.

Release managers: Post-deploy verification. "Did error rates spike after the last deployment?" for quick canary validation.

Dynatrace UI vs AI Agent

Both work. The AI agent is fastest during incidents when every second of navigation matters.

Combining with Other Observability MCP Servers

Dynatrace pairs well with other monitoring tools:

• PagerDuty + Dynatrace: Alert from PagerDuty, full context from Dynatrace in one conversation
• Slack + Dynatrace: Incident thread in Slack, detailed metrics from Dynatrace
• Prometheus + Dynatrace: Custom metrics from Prometheus, full-stack context from Dynatrace

Getting Started

Create your Dynatrace MCP server in under 60 seconds at DataFaucet. No API documentation needed. Works with any Dynatrace environment your browser can reach.