Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.honeycomb.io/llms.txt

Use this file to discover all available pages before exploring further.

​
Overview

Honeycomb Intelligence is a suite of AI-powered features built into Honeycomb that surfaces insights from your telemetry data so you can investigate faster, understand your systems more deeply, and resolve incidents with confidence.

​
What is Honeycomb Intelligence?

Honeycomb Intelligence is not a single feature—it is a set of capabilities that meet you where you are in your workflow. Some features are interactive, responding to your questions and guiding investigations in real time. Others work in the background, learning what normal looks like for your systems and alerting you when something meaningful changes. Use Honeycomb Intelligence to:
  • Investigate an incident interactively: Canvas is an AI-guided workspace inside Honeycomb that pairs natural language interaction with rich, interactive visualizations. Ask Canvas a question, explore query results, and share snapshots with teammates, all without leaving Honeycomb.
  • Troubleshoot production issues from your IDE: The Honeycomb MCP Server lets AI-powered IDEs and agents query your observability data directly using natural language. Run BubbleUp, detect outliers, and visualize traces without switching tools.
  • Get early warning of performance problems: Anomaly Detection (Early Access) learns what normal looks like for your services and automatically surfaces meaningful deviations before they affect your users. It reduces false positives and alert fatigue so your team focuses on real problems.
  • Build queries faster: Query Assistant is a text-to-query interface that helps you construct valid, runnable Honeycomb queries from a natural language description of what you want to find.
  • Create calculated fields faster: AI Assisted Calculated Fields helps you write valid calculated field expressions using natural language, drawing on your dataset schema to produce accurate results.

​
How Honeycomb Intelligence works

Honeycomb Intelligence is designed to fit into your existing workflow with minimal friction and full transparency into how it works.

​
AI models

Honeycomb Intelligence uses a combination of large language models (LLMs) and statistical machine learning models, depending on the task. Interactive features like Canvas, Query Assistant, and AI Assisted Calculated Fields use LLMs to interpret your input and generate responses grounded in your dataset schema and telemetry. Anomaly Detection uses statistical machine learning models rather than LLMs. It continuously builds a per-service baseline from your existing telemetry and surfaces deviations that are likely to indicate a real problem.

​
Data and model training

Honeycomb does not use your telemetry data to train foundation models. For details about model providers and data handling, visit Honeycomb AI Policies.

​
Data pipeline

Honeycomb Intelligence works with your existing telemetry data. You don’t need to reconfigure your instrumentation, set up a separate AI pipeline, or move data anywhere new to use them.

​
Team control

Honeycomb Intelligence is optional. Team Owners can enable or disable it at any time. When disabled, no Honeycomb Intelligence features are active, including features that surface insights passively. To learn how to enable or disable Honeycomb Intelligence for yout Team, visit Manage Team Behavior.