OTelBench is an open benchmark suite for evaluating AI agents and tools on real-world OpenTelemetry observability tasks. It provides datasets and scoring scripts that allow researchers and engineering teams to measure performance on instrumentation and tracing challenges.