worker-benchmarks

Benchmark worker performance by running the full suite to measure latency, throughput, memory usage, and failure rates across triggers, registry CRUD, agent selection, model cache, and concurrent workers.

Benchmarking collects latency, throughput, histogram, memory usage, and failure rate metrics to guide tuning and capacity planning for the worker system under real-world workloads.

Yes, you can run deterministic benchmarks via CLI commands or programmatically through TypeScript imports, defining specific targets, iterations, and metrics for each test run.

Compare baseline performance by executing comprehensive benchmark test runs before and after optimizations, using collected metrics to validate improvements in worker scenarios.

Yes, the benchmarking tool supports concurrent worker scenarios, evaluating how the agentic-flow worker system performs under real-world workloads across multiple simultaneous test runs.

Measure model cache and agent selection performance by applying the benchmarking suite to these specific targets, collecting latency and throughput metrics to guide tuning.