Dynatrace Portal

Connect Claude Code to the Dynatrace platform for real-time observability, application performance monitoring, and infrastructure health analysis through MCP.

When to Use This MCP Server

Connect this server when you need to:

• Query Dynatrace for real-time application performance metrics, error rates, and response time data while debugging performance issues in your code
• Investigate infrastructure health including host metrics, container performance, and service dependencies without opening the Dynatrace web interface
• Correlate code changes with performance impacts by analyzing deployment markers, error spikes, and metric trends from your development environment

Consider alternatives when:

• You use a different observability platform like Datadog, New Relic, or Grafana that has its own MCP integration
• You need to configure Dynatrace monitoring rules, alerting policies, or dashboard layouts, which are better handled through the Dynatrace web UI

Quick Start

Configuration

Copy
name: dynatrace-portal
type: mcp
category: devtools

Example Connection

Copy
claude mcp:connect dynatrace-portal

Available Tools

query-metrics:      Retrieve time-series metrics for hosts, services, or custom dimensions
get-problems:       List active and recent problems detected by Dynatrace Davis AI engine
get-entities:       Enumerate monitored entities including hosts, services, and processes
query-logs:         Search and filter log entries from Dynatrace log monitoring
get-traces:         Retrieve distributed traces for end-to-end request flow analysis

Core Concepts

Dynatrace MCP Integration Overview

Aspect	Details
Transport	stdio-based process via npx running @dynatrace-oss/dynatrace-mcp-server
Authentication	Dynatrace platform token with API access scopes for data retrieval
Data Model	Entity-metric model with automatic topology discovery and dependency mapping
AI Engine	Davis AI for automatic root cause analysis and anomaly detection
Query Language	Dynatrace Query Language (DQL) for flexible metric and log queries

Dynatrace Observability Architecture

Claude Code IDE
    |
    v
[MCP Client] --stdio--> [Dynatrace MCP Server]
                              |
                              v
                     [Dynatrace REST API v2]
                              |
                    +---------+---------+
                    |         |         |
                    v         v         v
              [Metrics   [Problems  [Logs &
               API]       API]      Traces]
                    \         |         /
                     [Dynatrace Platform]
                       Davis AI Engine

Configuration

Parameter	Type	Default	Description
DT_PLATFORM_TOKEN	string	required	Dynatrace platform access token with appropriate API scopes
DT_ENVIRONMENT	string	required	Dynatrace environment URL (e.g., https://abc12345.live.dynatrace.com)
default_timeframe	string	last2h	Default time range for metric and log queries
metric_resolution	string	1m	Default data point resolution for time-series metric queries
max_results	integer	100	Maximum number of results returned from entity and log queries

Best Practices

1. Scope Tokens to Minimum Required Permissions - Create a dedicated Dynatrace API token with only the scopes needed for your workflow. Common scopes include metrics.read, entities.read, logs.read, and problems.read.

2. Use Narrow Time Ranges for Faster Queries - Query metrics and logs with the smallest time range that covers your investigation window. Broad time ranges produce more data but are slower to retrieve and harder to analyze.

3. Start with Problems for Root Cause Analysis - When investigating issues, begin with get-problems to see what Davis AI has already detected. Davis provides root cause analysis and impact assessment that saves significant manual investigation time.

4. Correlate Deployments with Metrics - After deploying code changes, query metrics for the affected services to verify performance has not regressed. Compare pre-deployment and post-deployment time ranges to identify any impact.

5. Use Distributed Traces for Cross-Service Debugging - When a request traverses multiple services, use get-traces to follow the complete request path. Traces reveal which service in the chain is contributing latency or errors.

Common Issues

1. Token Authentication Fails with 401 - Verify the DT_PLATFORM_TOKEN has not expired and includes the required API scopes. Dynatrace tokens can be scoped to specific APIs, so a token with only metrics.read cannot access problems or logs.

2. Environment URL Format Incorrect - The DT_ENVIRONMENT URL must include the full environment ID (e.g., https://abc12345.live.dynatrace.com). Do not include trailing slashes or API path segments in the URL.

3. Metrics Query Returns No Data - The metric key may be incorrect or the specified time range may predate the metric's availability. Use get-entities to verify the monitored entity exists and has active metric collection.