---
title: "Honeycomb MCP: review, setup & limitations · Undominated.ai"
canonical: https://undominated.ai/mcp-servers/honeycomb/
description: "Explore production telemetry and save investigations in Honeycomb, with additional write scope for operational changes."
---

# Honeycomb MCP: review, setup & limitations · Undominated.ai

> Explore production telemetry and save investigations in Honeycomb, with additional write scope for operational changes.

[← Explore all mcp servers](/mcp-servers/)

OBSERVABILITY / Honeycomb

# Honeycomb MCP

Explore production telemetry and save investigations in Honeycomb, with additional write scope for operational changes.

 [See setup guidance ↓](#setup)[Original source ↗](https://docs.honeycomb.io/integrations/mcp/)

SOURCE REVIEW

 Reviewed 2026-10-14
 Evidence 3 linked sources
 Publisher Honeycomb
 Licence Not established
 [Read what was—and wasn’t—checked ↓](#review)

“It lets AI agents query, explore, and iterate on telemetry data just like you do in our UI.”

 [Honeycomb · upstream description ↗](https://docs.honeycomb.io/integrations/mcp/) Our analysis follows below.

01 / THE REASONING

## Why this made the selection.

 - Telemetry queries and investigation tools support iterative diagnosis from an agent.
- Explicit write scope separates basic reads from changes to Boards, Triggers, SLOs and other resources.
- Honeycomb’s overview says the MCP server exposes query, trace, board, and investigation tools, and that access starts limited until a write scope is granted.
- The configuration guide prefers OAuth and reserves API keys for agents that cannot log in interactively, with key creation limited to team owners.

### A good fit for

 - Investigate unexpected latency or failures in Honeycomb.
- Capture an approved investigation on a shared Board.
- Querying Honeycomb traces or boards from an agent that can complete OAuth
- An unattended agent only when a team owner has created an API key and the needed scopes

### Weigh up before choosing

 - Read access exposes telemetry to the connected client.
- Write tools require mcp:write; enabling that scope allows changes to alerting and investigation resources.
- The endpoint must match the US or EU region, and some environment access is role-restricted.
- Most tools are described as read-only. Write tools need the mcp:write scope granted during consent or key setup; the page does not say the default connection includes them.
- A tool-limiting header that is empty or unrecognized causes Honeycomb to reject requests that include it.
- No licence was stated on these pages, and none was inferred.
- This review did not execute the resource.
- The server was not called.
- A tool-list filter reduces what the client loads; it does not independently reduce the underlying API key authority.

02 / THE REVIEW RECORD

## What we actually inspected.

Source review has boundaries. A clear record is more useful than a “safe” badge.

### Material inspected

 - Honeycomb MCP overview, configuration guide and tool reference.
- https://docs.honeycomb.io/integrations/mcp/
- https://docs.honeycomb.io/integrations/mcp/configuration-guide
- https://docs.honeycomb.io/integrations/mcp/tools
- Round 2 reread of acquired publisher setup and permission evidence; see immutable wave review.

### Our findings

 - Most tools are read-only, but the server is not universally read-only.
- OAuth is the preferred interactive setup. Management API keys are an alternative for unattended clients, with explicit MCP and environment scopes.
- Honeycomb documents MCP as discrete tools for tasks such as running a query, fetching a trace, creating a Board, or starting a Canvas investigation, with access limited until permission is granted.
- Most tools are read-only. OAuth is preferred, and an API key is for agents that cannot use interactive login. Only team owners can generate that key.
- The tools page says Honeycomb MCP enforces per-tool rate limits, with a default of 50 calls per minute for most tools and lower limits for expensive or mutating ones.
- Honeycomb exposes a client tool-loading filter, and API-key setup requires a team owner plus the MCP read permission.

### Not established by this review

 - Server launch, authenticated tool calls, and release-to-source parity were not tested.
- The resource was not executed and its tools were not called.
- The server was not called. A source review is not a runtime or security certification.

The review applies to the material and revision named here. A newer upstream release can change its behavior.

03 / PUT IT TO WORK

## Connect a server deliberately.

[Upstream setup instructions ↗](https://docs.honeycomb.io/integrations/mcp/configuration-guide)

DOCUMENTED COMMAND

 claude mcp add honeycomb --transport http https://mcp.honeycomb.io/mcp Copy command ↗

Copying does not execute this command. It may retrieve a newer version than the reviewed source.

 - For a US account in Claude Code, use the command below; use the EU endpoint from the official setup guide for EU data.
- Finish OAuth and review requested read/write scopes.
- Use the client-specific configuration format; the JSON example is the documented Cursor shape for the US endpoint.

### Before you start

 - Honeycomb team access and telemetry data.
- An OAuth-capable client, or a suitably scoped management API key for unattended access.

### Documented configuration example

Use the client and authentication context described in the setup guidance. Configuration formats vary between hosts.

 {
 "mcpServers": {
 "honeycomb": {
 "url": "https://mcp.honeycomb.io/mcp"
 }
 }
} Copy configuration ↗

### Compatibility

Claude Code · Claude Desktop · Cursor · VS Code

### Transports

HTTP

### Remote OAuth or scoped API key

 - Queries telemetry and returns investigation results.
- With explicit write scope, creates/updates Boards, Triggers, SLOs, notification recipients and Canvas investigations.

### Cost model

Honeycomb plan and telemetry usage apply; the connected AI client is billed separately.

04 / FOLLOW THE EVIDENCE

## The source trail.

Our notes are separate from the original resource. Check upstream before adopting a new version.

 - [Honeycomb MCP setup and scope ↗](https://docs.honeycomb.io/integrations/mcp/) Checked 2026-10-14 https://docs.honeycomb.io/integrations/mcp/ Supports: summary, upstreamDescription, whySelected, bestFor, limitations, install, access, compatibility, transports, review
- [Honeycomb regional setup and authorization ↗](https://docs.honeycomb.io/integrations/mcp/configuration-guide) Checked 2026-10-14 https://docs.honeycomb.io/integrations/mcp/configuration-guide Supports: install, access, compatibility, transports, review
- [Honeycomb tool scope reference ↗](https://docs.honeycomb.io/integrations/mcp/tools) Checked 2026-10-14 https://docs.honeycomb.io/integrations/mcp/tools Supports: whySelected, limitations, access, review

KEEP COMPARING

## Other approaches to consider.

Related by category or shared topics. These are alternatives to inspect, not a measured quality order.

 [### Datadog MCP ↗ Query Datadog telemetry and operate permitted monitoring resources from an authenticated AI client.](/mcp-servers/datadog/)[### Jaeger Telemetry MCP Server ↗ In-tree Jaeger query service MCP server exposing OpenTelemetry distributed tracing queries over Streamable HTTP and SSE.](/mcp-servers/jaeger/)[### Logfire MCP ↗ Connects assistants to Logfire traces, metrics and observability management tools in the chosen region.](/mcp-servers/logfire/)

 [AI Tools ↗](/tools/)[Skills ↗](/skills/)[Agents ↗](/agents/)[MCP Servers ↗](/mcp-servers/)[Workflows ↗](/workflows/)

## Continue your investigation

 - [Inspect resource evidence](/tools/)
- [Choose a host](/agents/)
- [Inspect reusable workflows](/skills/)
- [Use Undominated data](/api/)
