Docs

Components

Metrics

• Counter: A cumulative metric representing a single incrementing counter whose value can be incremented or reset to 0 on restart.
• Gauge: A metric that represents a single numeric value whose value can go up or down arbitrarily.
• Histogram: Samples observations and counts the number of observations that fall into configurable buckets. It primarily measures things like request duration and response size.
• Summary: Similar to a histogram, but also provides total sums and counts of observed values, as well as configurable quantiles.

Metric Registry

The metric registry is the central place where all metrics are registered. This is an instance of the Registry class that stores a reference to a metric object.

Features of Metric Registry:

1. Collection: The registry collects metric data from a variety of sources. These metrics are typically instrumented within application code using the Prometheus client library.
2. Storage: The registry stores all registered metrics and their current values.
3. Scraping: The registry allows the Prometheus server to scrape data at regular intervals.
4. Exposure: The registry exposes metrics in a format that Prometheus can understand (usually plain text over HTTP).

Interaction with the Prometheus client library

The Prometheus client library provides the tools you need to instrument your application and generate metrics. These libraries typically include:

• Create metrics (counters, gauges, histograms, summary)
• Register metrics in registry
• Exposing metrics via HTTP endpoint