Grafana's Pyroscope 2.0 Makes Continuous Profiling Practical at Scale

"Grafana Labs has released Pyroscope 2.0, a ground-up rearchitecture of the open source continuous profiling database. The release, announced on 21 April 2026, addresses storage costs, query performance and operational complexity that had accumulated in the original design."

"Continuous profiling sits alongside metrics, logs and traces as the four pillars of an observability stack. Whilst metrics report issues such as high CPU usage, and traces show which service is the bottleneck, a profile goes further, identifying which function, on which line, is consuming the cycles."

"The most significant cost reduction comes from removing write-path replication. In v1, every profile was written three times. A single profile can be tens of megabytes, so that 3x amplification had a meaningful effect on storage bills. Pyroscope 2.0 writes each profile once to object storage."

"Pyroscope 2.0 applies the same architectural principles, adapted for the characteristics of profiling data: large payloads, significant symbolic information, and bursty query patterns. The second saving comes from data co-location: profiles from the same service are stored together, allowing symbolic information such as function names, source locations an"

Pyroscope 2.0 is a ground-up rearchitecture of the continuous profiling database, targeting storage costs, query performance, and operational complexity. Continuous profiling complements metrics, logs, and traces by identifying which function and line consume CPU cycles. Pyroscope 2.0 moves away from the earlier Cortex-based foundation and applies architectural principles similar to newer designs in other Grafana Labs observability components. The largest cost reduction comes from removing write-path replication, so each profile is written once to object storage instead of three times. Additional savings come from co-locating profiles from the same service to reuse symbolic information such as function names and source locations, improving efficiency for bursty query patterns.

Read at InfoQ