The current scripts/benchmark.py tests injection latency sequentially in a for loop. Since the new Rust daemon is highly concurrent, this doesn't test its true throughput. Create a new async_benchmark.py script using asyncio.gather and httpx.AsyncClient to bombard the Gateway with hundreds of simultaneous concurrent requests to measure maximum requests-per-second (RPS).
The current
scripts/benchmark.pytests injection latency sequentially in aforloop. Since the new Rust daemon is highly concurrent, this doesn't test its true throughput. Create a newasync_benchmark.pyscript usingasyncio.gatherandhttpx.AsyncClientto bombard the Gateway with hundreds of simultaneous concurrent requests to measure maximum requests-per-second (RPS).