Honeycomb
By Honeycomb.io
Honeycomb is an observability platform built around high-cardinality, high-dimensionality event data, allowing engineers to ask arbitrary questions about production system behavior rather than relying on predefined dashboards.
Definition
Honeycomb is an observability platform built around high-cardinality, high-dimensionality event data, allowing engineers to ask arbitrary questions about production system behavior rather than relying on predefined dashboards.
Overview
Honeycomb was founded by engineers who had worked on infrastructure at Parse and Facebook, and its design reflects a specific belief about modern software: production systems, especially those built on microservices, fail in ways that pre-built dashboards can't anticipate. Rather than aggregating data into metrics up front, Honeycomb ingests wide, structured events — one event per request, with dozens or hundreds of fields attached — and lets engineers query them interactively. This approach supports high-cardinality fields (like user ID or request ID) and high-dimensionality (many attributes per event) without losing the ability to drill down, which is the central argument Honeycomb has made in popularizing the modern definition of observability: the ability to ask new questions of a system without shipping new code. Its BubbleUp feature, for example, lets engineers select an anomalous cluster of events in a graph and automatically see which attributes correlate with it. Honeycomb supports OpenTelemetry as a primary way to get data in, positioning it within the broader shift toward vendor-neutral instrumentation. It's generally adopted by engineering teams working on complex, high-scale distributed systems who have outgrown dashboard-and-alert-based monitoring and need to debug novel failure modes in production, a workflow closely tied to distributed tracing.
Key Features
- Wide, structured event model supporting high-cardinality and high-dimensionality data
- Interactive querying for ad-hoc investigation without predefined dashboards
- BubbleUp feature for automatically surfacing attributes correlated with anomalies
- Native OpenTelemetry support for vendor-neutral instrumentation
- Distributed tracing visualizations tied directly to the same event data
- Service-level objective (SLO) tracking built on the same underlying events