---
title: "Metrics"
description: "Understand the metrics Middleware collects, how to query them in dashboards, and how data points are aggregated over time and across hosts."
section: "Metrics"
breadcrumb: "Metrics > Metrics Overview"
canonical: "https://docs.middleware.io/features/metrics"
---
# Metrics

Metrics are the numbers behind every dashboard widget and alert: CPU and memory from your hosts, request rates and latencies from your services, and anything your integrations or custom instrumentation report. Middleware stores them in the OpenTelemetry model, so each metric has a type, such as gauge, sum or histogram, that decides how it can be queried. These pages explain the metric types, how to query them with the visual query builder, and exactly how raw data points are combined into the lines and numbers you see.

## Features

- [Types of Metrics](/features/metrics/types-of-metrics) — Gauges, sums, histograms and summaries, and how each is queried
- [Querying Metrics](/features/metrics/querying-metrics) — Pick, filter, aggregate and transform metrics in a dashboard widget
- [Understanding Aggregations](/features/metrics/understanding-aggregations) — How time aggregation, space aggregation and reducers turn data points into values

## Related Features

- [Dashboards](/features/dashboards) — Build widgets over your metrics with the visual query builder
- [Alerts](/features/alerts) — Create alert rules and service level objectives
- [Infrastructure](/features/infrastructure) — Hosts, Containers, Kubernetes etc.
- [APM](/features/apm) — Traces, services, profiles and endpoints