Prometheus Cheat Sheet

Updated 2026-04-28

Next Topic: Release Management Cheat Sheet

Prometheus is an open-source monitoring and alerting toolkit designed for reliability and scalability in cloud-native and containerized environments. Born at SoundCloud in 2012 and now a CNCF graduated project, Prometheus operates on a pull-based model, scraping metrics from instrumented targets at regular intervals and storing them in an efficient time-series database (TSDB). What makes Prometheus particularly powerful is PromQL, its functional query language that enables real-time aggregation and analysis—paired with a flexible alerting system that routes notifications through Alertmanager. With Prometheus 3.0 (November 2024), the project added UTF-8 metric names, native OTLP ingestion, Remote Write 2.0, and a redesigned UI; keep in mind that Prometheus is metrics-first, not logs or traces, and its dimensional label model demands careful attention to cardinality to avoid performance degradation.

What This Cheat Sheet Covers

This topic spans 26 focused tables and 201 indexed concepts. Below is a complete table-by-table outline of this topic, spanning foundational concepts through advanced details.

Table 1: Metric TypesTable 2: PromQL Selector TypesTable 3: PromQL Aggregation OperatorsTable 4: PromQL Rate and Delta FunctionsTable 5: PromQL Time, Offset, and Subquery ModifiersTable 6: PromQL Over-Time FunctionsTable 7: PromQL Label Manipulation FunctionsTable 8: PromQL Histogram FunctionsTable 9: PromQL Math FunctionsTable 10: PromQL Arithmetic and Comparison OperatorsTable 11: PromQL Logical and Set OperatorsTable 12: PromQL Vector Matching ModifiersTable 13: PromQL Utility and Datetime FunctionsTable 14: Prometheus Configuration SectionsTable 15: Service Discovery MechanismsTable 16: Relabeling ActionsTable 17: Alerting Rule ComponentsTable 18: Recording Rule StructureTable 19: Alertmanager Configuration ConceptsTable 20: Common Prometheus ExportersTable 21: Prometheus HTTP API EndpointsTable 22: Storage and Retention ConfigurationTable 23: Prometheus Agent ModeTable 24: Label Best Practices and CardinalityTable 25: Client Libraries for InstrumentationTable 26: promtool Commands

Table 1: Metric Types

Type	Example	Description
Counter	`http_requests_total{method="GET"} 1523`	• Cumulative value that only increases or resets on restart • use `rate()` or `increase()` to derive per-second rates or totals.
Gauge	`node_memory_active_bytes 4.2e+09`	• Instantaneous value that can go up or down • represents current state like memory usage, queue depth, or temperature.
Histogram	`http_request_duration_seconds_bucket{le="0.5"} 124` `http_request_duration_seconds_sum 53.2` `http_request_duration_seconds_count 144`	• Samples observations into configurable buckets with `_sum` and `_count` • use `histogram_quantile()` for server-side approximate percentiles.

Table 1: Metric Types

Type	Example	Description
Counter	`http_requests_total{method="GET"} 1523`	• Cumulative value that only increases or resets on restart • use `rate()` or `increase()` to derive per-second rates or totals.
Gauge	`node_memory_active_bytes 4.2e+09`	• Instantaneous value that can go up or down • represents current state like memory usage, queue depth, or temperature.
Histogram	`http_request_duration_seconds_bucket{le="0.5"} 124` `http_request_duration_seconds_sum 53.2` `http_request_duration_seconds_count 144`	• Samples observations into configurable buckets with `_sum` and `_count` • use `histogram_quantile()` for server-side approximate percentiles.