Can AI agents automate Honeycomb?

Honeycomb observability is automatable today through the official hosted MCP. This record is Honeycomb query, investigation, Boards, Triggers, SLOs, and related team resources as exposed by https://mcp.honeycomb.io/mcp (US) and https://mcp.eu1.honeycomb.io/mcp (EU). It is not Datadog, New Relic, Grafana, or Dynatrace. The archived honeycombio/honeycomb-mcp GitHub repository is not the current product. A Team Owner must enable Honeycomb Intelligence before MCP works. OAuth is the documented default. Unattended agents may use a Management API key that a Team Owner creates with Model Context Protocol and Environments scopes, sent as Authorization: Bearer KEY_ID:SECRET. Read is required for all MCP operations; write (mcp:write) is required to create or update Boards, Triggers, SLOs, recipients, and Canvas investigations. Documented tools include workspace discovery, run_query, traces, BubbleUp, Boards, Triggers, SLOs, recipients, and Canvas. Per-tool rate limits apply. get_service_map is Enterprise. The REST APIs at api.honeycomb.io and api.eu1.honeycomb.io manage datasets, query specs, Boards, Triggers, SLOs, Environments, and keys. Configuration and Ingest keys use the X-Honeycomb-Team header. Management keys use Authorization: Bearer. Creating a Query spec does not run it. The Query Data API that returns results is Enterprise-only, with a 7-day window and 10 requests per minute. Official send-data libraries include Libhoney and OpenTelemetry paths. honeymarker is a marker CRUD CLI, not a general Honeycomb CLI. Trigger and SLO burn alerts can POST JSON to webhooks. iPaaS catalogs were not opened. Computer-use viability is unassessed. MCP scores 8 and is the best path because it can run queries without the Enterprise Query Data API. The REST API scores 7 because query results stay Enterprise- gated.

Best path today: mcp · Overall automatability: 8/10