Honeycomb

What it is

Honeycomb is a commercial observability platform founded in 2016 by Charity Majors and Christine Yen (both former Parse/Facebook engineers), built around ingesting and querying arbitrarily wide, structured events rather than pre-aggregated metrics. Its core technical bet is that high-cardinality, high-dimensionality data — unique request IDs, user IDs, build versions, and dozens of other attributes per event — should be queryable in real time without needing to predict in advance which dimensions matter, which is what its columnar storage engine is built to support.

Majors and Yen are also credited with popularizing the modern definition of “observability” as distinct from traditional monitoring, and the company has continued pushing that framing forward with what it calls “Observability 2.0”: the idea that a single stream of wide structured events (rather than separate metrics, logs, and traces pipelines) should be the primary source of truth, with traces, logs, and metrics-like aggregates all derivable from that one dataset. In 2026 the company expanded into first-class metrics support and deeper integrations for monitoring AI-assisted development workflows, extending that model beyond its original trace-and-event focus.

Why teams choose it

• Genuinely fast high-cardinality queries. Honeycomb’s query engine is built specifically to slice by arbitrary, high-cardinality fields (like individual user or request IDs) without pre-defined indexes, which is a real differentiator versus metrics-first tools that struggle with cardinality explosions.
• BubbleUp and similar exploratory tooling. Features that automatically surface which dimensions correlate with an anomaly are genuinely useful for root-causing issues you didn’t know to look for, rather than requiring a pre-built dashboard.
• Instrumentation discipline required. Getting full value requires embracing wide-event instrumentation (rich, high-dimension events) rather than traditional narrow metrics/logs, which is a real adoption cost for teams with existing instrumentation invested elsewhere.
• Smaller ecosystem than the big commercial players. Compared to Datadog or New Relic, Honeycomb has a narrower first-party feature surface (less emphasis on infrastructure monitoring, RUM, or security), so larger organizations often pair it with other tools rather than using it as a single pane of glass.

Pricing model

Honeycomb uses usage-based pricing keyed to event volume ingested per month, with a free tier for small-scale or evaluation use and paid tiers (Pro and Enterprise) scaling with event throughput and retention needs. It has recently introduced separate consumption-based pricing for its newer Metrics product, priced by time-series volume rather than event count, reflecting the fact that metrics and wide-event tracing are billed somewhat differently within the platform.

Alternatives

Grafana Tempo and Jaeger are open-source tracing backends that require self-hosting but avoid usage-based billing entirely, at the cost of Honeycomb’s high-cardinality query experience. Datadog offers similar tracing and event-analysis capability bundled with a much broader all-in-one platform. Splunk Observability Cloud is another commercial alternative with a comparable focus on high-cardinality, real-time analysis of trace and event data.