Hey everyone,
We're running a gRPC service on Cloud Run and keep hitting this error after a few minutes under load: Error: 8 RESOURCE_EXHAUSTED: Bandwidth exhausted or memory limit exceeded Our setup: - 50 requests per second - Each request takes 3-6 seconds to respond - Memory: 512 MiB - CPU: 1 - Max concurrency: 20 per instance - Max instances: 100 The service works fine for short bursts, but once we hit sustained traffic for a few minutes, multiple instances start throwing this error. I found that Cloud Run has a 600 Mbps bandwidth limit per instance. Our gRPC responses might be pretty large, so I'm thinking we're hitting that limit. Questions: 1. Is this definitely a bandwidth issue, or could it be something else? 2. Should we lower the concurrency to spread traffic across more instances? 3. Will bumping up memory/CPU help with bandwidth, or are those limits fixed? 4. Any other config changes we should try? We're a bit stuck here - the service works great until it hits this wall. Any advice would be awesome! Thanks! -- You received this message because you are subscribed to the Google Groups "grpc.io" group. To unsubscribe from this group and stop receiving emails from it, send an email to grpc-io+unsubscr...@googlegroups.com. To view this discussion visit https://groups.google.com/d/msgid/grpc-io/ea5c82c3-a6b3-4328-8ea8-c554d3039e2bn%40googlegroups.com.
