Honeycomb: Migrating to Time Series Metrics

Honeycomb is an observability platform used by software engineers to debug complex, high-volume production systems. It originally stored all telemetry data (traces, logs, events) in a "wide events" format that preserves rich, high-cardinality data — meaning it kept every unique attribute and value (e.g., user_id, request_id, container_id) intact for deep, ad-hoc investigation. The problem: storing every raw event this way is expensive and slow when you only need aggregates or long-term trends. Time Series Metrics are a newer, cheaper storage tier that pre-aggregates data into numeric counters and histograms, discarding most individual event details. This migration means users will shift some of their monitoring to a more traditional metrics-like system within Honeycomb. Companies like Honeycomb (founded 2016) compete with Datadog, New Relic, and Grafana in the "observability" space, where balancing data richness vs. cost is a constant tension. Understanding this migration helps engineers decide which data belongs in which tier and why.