Overview
Note: This self-hosted MCP server is deprecated. A single server process runs on your own computer and communicates with clients over STDIO; it is not authenticated. It enables LLMs like Claude to directly analyze and query Honeycomb observability data across multiple environments. It requires Node.js 18+ and a Honeycomb API key with full permissions (analytics queries, read access for SLOs and Triggers, and environment-level dataset operations). Configuration uses an MCP config that launches a Node process pointing to the built index.mjs, with environment variables such as HONEYCOMB_API_KEY and per-environment keys (HONEYCOMB_ENV_PROD_API_KEY, HONEYCOMB_ENV_STAGING_API_KEY) for multi-env setups. EU customers can set HONEYCOMB_API_ENDPOINT to use the EU instance. The MCP server implements TTL-based caching for non-query API calls to improve performance and reduce API usage. It supports querying across environments, analytics queries with various calculations, monitoring SLOs, analyzing columns, viewing Triggers, and retrieving dataset metadata and schema. Client compatibility includes Claude Desktop, Claude Code, Cursor, Windsurf, Goose, and others. This self-hosted option is intended for Honeycomb Enterprise customers, with hosted MCP recommended going forward.
Features
Query Honeycomb datasets across multiple environments
Access and query Honeycomb datasets across environments via the MCP server.
Run analytics queries with support for multiple calculation types, breakdowns, filters, and time-based analysis
Perform analytics using operations like COUNT, AVG, P95, etc., with breakdowns, filters, and time ranges.
Monitor SLOs and their status
Observe SLOs and their status (Enterprise only).
Analyze columns and data patterns
Analyze column data and detect patterns and insights.
View and analyze Triggers
Inspect and analyze trigger configurations and behavior.
Access dataset metadata and schema information
Retrieve dataset metadata, column details, and schema information.
Optimized performance with TTL-based caching for non-query API calls
TTL-based caching improves performance and reduces Honeycomb API usage.
Who Is This For?
- Enterprise customers:Enable internal teams to securely query and analyze data across environments using an on-prem MCP server.
- LLM integrators:Leverage LLMs to analyze Honeycomb data locally across environments.
- Data scientists:Data scientists build analytics workflows across datasets with cross-environment access.
