Replies: 2 comments 8 replies
-
You need to tune concurrency at both ends. Right now, your RPC server has a default pool size (I guess) of 30 threads, but you try to make ~1k concurrent requests from the
It depends on what these clients are doing; do you have any monitoring set up? |
Beta Was this translation helpful? Give feedback.
-
Hi @palkan, We are using client-side load balancing via ANYCABLE_RPC_HOST. This env var points to a headless service. About health checks, do you mean http-health-port and readinessProbe config ? With my config, the ResourceExhausted errors are due to the headless service not working or rpc pods getting down, is that it? Thanks |
Beta Was this translation helpful? Give feedback.
-
i use custom setting when start anycable server to try speed up server
bundle exec anycable --log-file=log/anycable.log --log-level debug --server-command "\anycable-go --host=0.0.0.0 --port=59901 --rpc_concurrency=1020 --hub_gopool_size=640"
but i will show error like that
how to fix it ?
and what is correctly config setting if I want support 10k websocket client request, every response can in 6 second to finish.
Beta Was this translation helpful? Give feedback.
All reactions