Hello , we have a performance problem with nakama and cockroachdb.
We have 10 thousand concurrent users , after a few hours there are more than 30 thousand ports in wait_time state.
we get this message in nakama:
{“level”:“error”,“ts”:“2023-12-08T14:42:18.971+0330”,“caller”:“server/core_friend.go:135”,“msg”:“Error retrieving friends.”,“error”:“timeout: context canceled”}
We can’t advise on custom deployments because we don’t have access to the infrastructure and there are too many different variants. If you need an easy auto-scalable solution, use Heroic Cloud which will also fund the development of the Open-Source technology.