metrics-queries

Querying OpenTelemetry metrics in Honeycomb requires understanding correct syntax for gauges, counters, and histograms, plus overriding default temporal aggregations to calculate rates and utilization percentages accurately.

You calculate the 99th percentile of http.server.duration by querying the histogram metric and grouping by service.name to get the exact percentile breakdown for your services over the last hour.

Honeycomb metrics queries return different results than event datasets because metrics use temporal aggregation by default, which automatically summarizes data points over time windows before applying your query operations.

You can discover available metrics and their associated attributes within a dataset by exploring the schema, which allows you to identify what OpenTelemetry instruments are available to query.

You override default temporal aggregation for rate calculations by applying specific query operations that adjust how metrics data is summarized, allowing you to compute custom rates and utilization percentages instead of relying on default rollups.

The best way to query histogram metrics in Honeycomb is to use the correct query structure for percentile calculations, ensuring you understand how temporal aggregation affects the distribution data before applying your final aggregations.