We are excited to announce that Dash0 now supports Exponential Histograms — the next-generation histogram format introduced by OpenTelemetry.
Dash0 has already been ingesting and storing exponential histograms from OpenTelemetry sources, but until now these metrics could only be listed in the UI. With this update, exponential histograms can be queried and visualized, allowing you to explore, graph, and analyze these metrics alongside regular histograms.
Unlike regular histograms with fixed bucket boundaries, exponential histograms use dynamically scaled buckets that automatically adjust to the magnitude of recorded values. This allows Dash0 to:
- Efficiently represent metrics with high dynamic ranges (for example, latency from microseconds to seconds).
- Maintain higher precision without excessive memory usage.
- Provide smoother and more accurate visualization for metrics such as latency, request duration, or payload sizes.
Dash0 ingests and stores OpenTelemetry exponential histograms in their native form, translating them into native histogram representation for querying and visualization in the UI. This ensures full compatibility with OpenTelemetry data while preserving the precision and scale of the original metrics.
This update allows you to use any of the standard Prometheus histogram functions — such as histogram_quantile() — with full precision and scale preserved from the original OpenTelemetry data.