It seems likely that Phoenix can run in production with minimal resources; internally, Arize uses 2GB RAM and a single CPU for hosted Phoenix. For production, you can start with 2 vCPUs and 3GB RAM, but actual requirements depend on trace volume, user concurrency, and retention settings. Phoenix is stateless and supports horizontal scaling, so you can deploy multiple containers behind a single Postgres database as your workload grows. Monitor memory and disk usage under expected load and scale up or out as needed. There are no official load test benchmarks published, so start small and adjust based on observed performance and ingestion rates.For production reliability, enable batch processing, use gRPC transport, and ensure automated Postgres backups. Memory and disk needs will increase with higher ingestion rates, more unique labels, and longer retention. Regularly monitor resource utilization and database size to plan for scaling. Using Postgres (not SQLite) is recommended for production deployments on Google Cloud. Would you like more detail on scaling strategies or configuration specifics?Sources: - Slack: Internal resource sizing and scaling - Production Guide: Scaling and resource planning - Production Guide (GitHub)
馃挕聽Hint: Mention RunLLM in the thread for followups.
