OpenTelemetry Observability Standard

OpenTelemetry (OTel) is a free and open-source observability framework for cloud-native software, providing a unified standard for generating, collecting, and exporting telemetry data (traces, metrics, and logs). OpenTelemetry was formed in 2019 by the merger of two earlier projects: OpenTracing (from CNCF, launched 2016) and OpenCensus (from Google, launched 2015). It is the second most active project in the CNCF (Cloud Native Computing Foundation), after Kubernetes, and is backed by major companies including Google, Microsoft, AWS, Datadog, and Dynatrace. Key features: traces: distributed tracing tracks requests as they flow through services in a distributed system. A trace consists of spans (units of work), each with a name, start time, duration, attributes, events, and status. Spans are linked in a parent-child hierarchy to form a trace tree. Context propagation: trace context is propagated across service boundaries via HTTP headers (W3C Trace Context) or message queue metadata, enabling end-to-end tracing across microservices. Metrics: numerical measurements collected at intervals, including counters (monotonically increasing), gauges (can go up or down), and histograms (distribution of values). Metrics can be collected via push or pull models. Logs: structured log records correlated with traces and metrics. The logs API and SDK are designed to bridge existing logging frameworks (Java Log4j, Python logging, Go zap) into the OpenTelemetry ecosystem. Specification: this repository contains the formal specification that defines the API, SDK, and data model requirements. Language SDKs: official implementations in Java, Python, Go, JavaScript/TypeScript, C++, Rust, .NET, Ruby, PHP, Swift, and Erlang. Collector: a vendor-neutral proxy that receives, processes, and exports telemetry data to multiple backends (Prometheus, Jaeger, Zipkin, Datadog). Apache-2.0.