MCP

The Dash0 MCP (Model Context Protocol) Server is designed to facilitate advanced interactions between AI assistants and your observability data.

Setup

Check the AI category in our integration hub for instructions on how to add the Dash0 MCP server to different AI assistants like Claude, Cline and Cursor.

Architecture

The Dash0 MCP server is fully hosted on our cloud infrastructure, in each of the availability regions, so you don't have to run anything extra on your side.

It uses the StreamableHttp transport, which is compatible with most modern MCP client implementations.

Capabilities

The Dash0 MCP server exposes capabilities that allow AI agents to navigate and consume your telemetry data in a similar way to how human users would do in the UI.

The available tools expose Dash0 features like:

  • Triage of error Logs and Spans
  • Catalog of services, operations and metrics
  • Detailed information about Failed Checks
  • The Query Builder and full PromQL support, including synthetic metrics for logs and spans

Examples

These are just a few examples of questions and tasks that can be easily solved by AI agents while leveraging the capabilities of Dash0's MCP server:

  • Investigate the root cause of a failed check
  • Write a post-morten for an incident from last week
  • Identify performance bottlenecks in a stack of services
  • Estimate the monthly cloud infrastructure costs
The images show an example where Claude is able to identify the root cause of an incident in the OpenTelemetry Demo dataset by employing tools from the Dash0 MCP server.

Last updated: June 13, 2025

Previous

API

Next

Integrations