bmad-observability-agent

Dynatrace Setup

Guide for configuring dtctl and the Dynatrace MCP server for use with the O11y Engineer.

dtctl Setup

dtctl is a kubectl-style CLI for managing Dynatrace configuration as code.

Installation

# Install via npm
npm install -g @dynatrace-oss/dtctl

# Or via brew
brew install dynatrace-oss/tap/dtctl

Configure a Context

# Set up a Dynatrace environment context
dtctl config set-context production \
  --url "https://abc12345.live.dynatrace.com" \
  --token "dt0s16.YOUR_API_TOKEN"

# Switch to the context
dtctl config use-context production

# Verify connection
dtctl get dashboards

Required Token Permissions

Scope	Purpose
ReadConfig	Read dashboards, settings
WriteConfig	Create/update dashboards, settings
entities.read	Read entity data
metrics.read	Read metrics
logs.read	Read log data
openpipeline.events	Read events
automation.workflows.read	Read workflows
automation.workflows.write	Create workflows
slo.read	Read SLOs
slo.write	Create SLOs
syntheticMonitoring.read	Read synthetic monitors
syntheticMonitoring.write	Create synthetic monitors

Project Structure

project/
  dynatrace/
    dashboards/
      service-overview.yaml
    notebooks/
      incident-analysis.yaml
    workflows/
      auto-remediation.yaml
    slos/
      service-slos.yaml

Deploy

# Validate
dtctl apply -f dynatrace/ --dry-run

# Deploy all
dtctl apply -f dynatrace/

# Deploy specific
dtctl apply -f dynatrace/dashboards/service-overview.yaml

---

Dynatrace MCP Server Setup

The Dynatrace MCP server enables AI-driven discovery and context-aware DQL generation.

Configuration

Add to your AI assistant’s MCP configuration:

=== “Claude Code”

Add to `~/.claude/settings.json` or project `.claude/settings.json`:

```json
{
  "mcpServers": {
    "dynatrace": {
      "command": "npx",
      "args": ["-y", "@dynatrace-oss/dynatrace-mcp-server"],
      "env": {
        "DT_ENVIRONMENT": "https://abc12345.apps.dynatrace.com",
        "DT_PLATFORM_TOKEN": "dt0s16.YOUR_TOKEN"
      }
    }
  }
}
```

=== “Cursor”

Add to `.cursor/settings.json`:

```json
{
  "mcpServers": {
    "dynatrace": {
      "command": "npx",
      "args": ["-y", "@dynatrace-oss/dynatrace-mcp-server"],
      "env": {
        "DT_ENVIRONMENT": "https://abc12345.apps.dynatrace.com",
        "DT_PLATFORM_TOKEN": "dt0s16.YOUR_TOKEN"
      }
    }
  }
}
```

=== “Windsurf”

Add to `.windsurfrules/mcp.json`:

```json
{
  "mcpServers": {
    "dynatrace": {
      "command": "npx",
      "args": ["-y", "@dynatrace-oss/dynatrace-mcp-server"],
      "env": {
        "DT_ENVIRONMENT": "https://abc12345.apps.dynatrace.com",
        "DT_PLATFORM_TOKEN": "dt0s16.YOUR_TOKEN"
      }
    }
  }
}
```

Verify MCP Connection

After configuring, restart your AI assistant and verify the MCP connection. If MCP is available, you’ll see:

Dynatrace MCP server is configured and connected.
Available capabilities:
- Execute DQL queries
- Discover services and entities
- Analyze logs, metrics, and traces
- Create dashboards and notebooks

MCP-Powered Commands

Menu Code	Purpose
PD	Project Discovery — find services, metrics in Dynatrace
DB	Dashboard Builder — build dashboard from real metrics
SW	Suggest Workflows — get AI-suggested automations
DN	Diagnostic Notebook — build troubleshooting notebook

Required Token Scopes for MCP

Scope	Purpose
storage:logs:read	Query logs
storage:spans:read	Query spans
storage:metrics:read	Query metrics
storage:entities:read	Discover entities
storage:events:read	Query events
automation:workflows:read	Read workflows

---

Using the O11y Engineer Command

Run SD (Setup Dynatrace via /o11y-engineer) to get an interactive guide through the complete setup process:

SD

This covers:

1. dtctl installation and context configuration
2. Token creation with correct scopes
3. MCP server configuration for your IDE
4. Connection verification
5. First dashboard or notebook creation

This site is open source. Improve this page.