Glossary · Observability
What Is PromQL?
Written by
Dilip Namdev
Published
May 2026
Reading time
5 min read
PromQL is the language most teams use to turn raw metrics into rates, percentiles and alerts.
PromQL is the Prometheus Query Language, used to select and aggregate time-series metrics. It underpins most metric dashboards and alerts and is widely used to query OpenTelemetry metrics stored in Prometheus-compatible backends.
What PromQL does
PromQL selects time series by metric name and labels, then applies functions and aggregations. Common patterns include taking the per-second rate of a counter, computing a percentile from a histogram, and aggregating across labels. It is a functional, expression-based language built for time-series data.
Why it matters
PromQL became the de facto standard for querying metrics, so knowing it transfers across Prometheus, Grafana and many OpenTelemetry-compatible backends. When you see a dashboard panel or an alert rule, it is very often a PromQL expression underneath, which is why the how-to dashboards on this site express metric panels as PromQL-style queries.
How it fits Opstral
Opstral works with Prometheus-compatible metrics, so teams keep the PromQL knowledge and dashboards they already have while Telemetry Ops and Sentinel AI add autonomous resolution on top.
Frequently asked questions
Is PromQL only for Prometheus?
It originated with Prometheus but is supported by many compatible backends, so PromQL skills transfer across much of the metrics ecosystem.
How do I compute a percentile in PromQL?
With the histogram_quantile function over a histogram metric, for example the 99th percentile of request duration.