It could be possible that Phoenix has some internal load testing scripts and limited public discussion of load testing, but there do not appear to be published, official performance test results or benchmarks for traces per second. There is a merged pull request for a load testing script, but no documentation of results or throughput numbers is provided in the available sources. You may need to run your own benchmarks using the provided scripts or reach out to the Phoenix team for more details on production performance expectations (test: Add load testing script).Though I'm not entirely certain, it might be that the Phoenix team has not published detailed load test results, and users are encouraged to start with minimal resources and scale up as needed based on their own workloads (Slack discussion). Would you like more detail on how to run your own load tests or interpret the available scripts?Sources:
- test: Add load testing script
- Slack discussion
馃挕聽Hint: Mention RunLLM in the thread for followups.
hi Andrei L. we do not currently have official benchmarking scripts but this depends on a number of factors, such as database performance and the topology of your input traces