instrumentation library
also: telemetry library, observability library, monitoring library
A software library that collects and exports telemetry data such as metrics, traces, and logs from an application to monitoring systems. It instruments code by injecting measurement points that observe runtime behavior without modifying business logic.
An instrumentation library provides APIs and components to gather observability data from running applications. It captures metrics (counters, gauges, histograms), distributed traces (request paths across services), and structured logs that help operators understand system behavior, diagnose problems, and optimize performance.
Common examples include the OpenTelemetry libraries (available for most languages), Prometheus client libraries, and Jaeger libraries. For instance, a Python web application might use the OpenTelemetry Flask instrumentation to automatically track HTTP request duration, status codes, and error rates without adding instrumentation code to each route handler.
Instrumentation libraries typically work by wrapping or patching existing APIs—such as database drivers, HTTP clients, or web frameworks—so telemetry is collected transparently. The collected data is then exported to backends like Prometheus, Grafana Loki, or cloud monitoring services for analysis and alerting.