Introduction

Query language for Prometheus, used to query metrics from time series data.
Prometheus stores metrics as time series identified by metric name and optional key-value labels.

Metric Types

Type	Description	Example
Counter	Monotonically increasing value (e.g., requests total)	`http_requests_total`
Gauge	Can go up and down (e.g., current memory usage)	`memory_usage_bytes`
Histogram	Buckets of observations (e.g., request durations)	`http_request_duration_seconds_bucket`
Summary	Quantiles over observations (e.g., p95 latency)	`http_request_duration_seconds`

Queries

Basic queries

// Get latest value of a metric
http_requests_total

// Filter by label 
http_requests_total{method="GET", status="200"}

//Filter by regex
http_requests_total{job=~"frontend|backend"}

Aggregation Operators

sum() — sum over a label

sum(http_requests_total)        # total requests
sum(http_requests_total) by (job)   # sum per job

avg() — average over time series

avg(http_requests_total) by (instance)

max(), min(), count(), stddev(), stdvar() — other aggregations