• DATA DISCOVERY
  • QUERYING
  • ROOT CAUSE
  • HUMAN REVIEW

How to query telemetry and troubleshoot incidents with your AI

Connect Honeycomb to your agent to turn natural language into precise observability queries. Once connected, your AI handles the heavy lifting of data exploration.

Ask AI about this page

Short answer

How can I use my AI to troubleshoot Honeycomb telemetry?

Your agent uses the Honeycomb MCP to list datasets, check available columns, and execute queries to find anomalies. You get specific query results that highlight exactly where a system is failing.

Where every issue lands.

Incident investigation outcomes

The AI moves through the observability lifecycle to narrow down your search.

DATA DISCOVERY

Map your telemetry

The AI identifies which datasets and columns are available for investigation. You get a clear view of the telemetry schema without manual searching.

QUERYING

Run targeted searches

Your agent translates your questions into functional queries. It pulls the exact data points needed to validate a hypothesis.

ROOT CAUSE

Isolate the failure

By analyzing query results, the AI identifies patterns or outliers. This points you directly to the broken component or service.

HUMAN REVIEW

Verify the findings

The AI presents the raw data and query logic. You make the final call on whether the identified cause matches the real-world incident.

The workflow

What your AI does when an incident occurs.

The agent acts as a bridge between your questions and the Honeycomb data lake.

  1. Survey the landscape

    The agent first finds the relevant telemetry by looking through your available data sources.

    list_datasets

  2. Inspect the schema

    To build a valid query, the agent checks which attributes and dimensions exist within the selected dataset.

    list_dataset_columns

  3. Execute the search

    The agent runs the specific query needed to isolate the error or latency spike.

    run_query

  4. Analyze the data

    The agent pulls the specific rows or aggregations required to answer your question.

    get_query_result

Try it

Copy these to start.

Use these prompts to kick off an investigation.

Starting points

Swap out the dataset names and error types for your specific environment.

Honeycomb Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_ezx3V1wqKIj03Mpp0VSR1HqDgg5PggVT54nkjF06/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — Honeycomb capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "honeycomb-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_ezx3V1wqKIj03Mpp0VSR1HqDgg5PggVT54nkjF06/mcp"
    }
  }
}
Copy into chat04

  • List all my Honeycomb datasets so I can see what telemetry is available.

  • What columns are available in the 'production_web_logs' dataset?

  • Run a query in 'api_gateway_metrics' to find all requests with a 500 status code in the last hour.

  • Get the results for the query that shows p99 latency spikes for the 'auth-service' service.

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect Honeycomb once, then ask.

One connection links your agent to your Honeycomb account. Your credentials stay encrypted and secure on our side.

Connect Honeycomb to your AI

FAQ

How the Honeycomb Connector behaves

  • 01

    Can the AI create new datasets in Honeycomb?

    No. The agent can only read existing data and run queries.

  • 02

    Can the AI delete my telemetry data?

    No. This MCP is read-only and cannot modify or delete any data in your Honeycomb account.

  • 03

    Does the AI write the query for me?

    Yes. The agent uses the available columns to construct and run queries based on your natural language requests.

  • 04

    Can the AI fix the code causing the error?

    No. The AI can find the error in the telemetry, but you must perform the actual code fix in your IDE.

  • 05

    How does the AI know which columns to use?

    The agent inspects the dataset schema to ensure the query uses valid column names.

  • More questions about Honeycomb? The Connector page answers them. See everything the Honeycomb Connector can do

Connect Honeycomb to Claude, Cursor, ChatGPT & more