Metrics in Prometheus provide quantitative data about system performance, health, and usage over time.

Prometheus collects metrics by scraping endpoints that expose data in a specific format over HTTP.

A metric is a time-series data point that represents a specific aspect of a system’s performance, such as CPU usage or memory consumption.

Labels are key-value pairs that provide additional information about the metric, such as the instance or job name.

A timestamp records the time at which a metric value was collected, ensuring that time-series data is accurate and can be plotted over time.

Histograms are used to collect data on the distribution of values, such as request durations or response sizes, and allow for calculating quantiles.

A counter is a metric that only increases over time, representing quantities like the number of requests processed or errors encountered.

Prometheus metrics are structured as a series of time-stamped data points that are organized by metric name and labels.

The Prometheus Pushgateway is used to allow short-lived jobs to push metrics to Prometheus, as it typically relies on pull-based collection.