1500 is probably fine 3000 is probably fine
One place you may see issues is on reconnect during a bounce - make sure you have your auth caches enabled and consider dropping the bcrypt rounds (defaults to 2^10, which takes some non-trivial cpu to hash the password supplied on each new connection, which you may notice when you restart). > On May 26, 2020, at 7:59 AM, Rahul Reddy <rahulreddy1...@gmail.com> wrote: > > > Hello, > > Our service code is written in goland uses gocql to connect todb and we have > 30 node cluster. And we have service connections coming from kubernettes pods > when we scaleup pods to 500 for each pod we are getting connections 1500 per > each node. We are not seeing any issues now but wanted to know if there is > limit on gocql to number of connections it can reach before erroring out --------------------------------------------------------------------- To unsubscribe, e-mail: user-unsubscr...@cassandra.apache.org For additional commands, e-mail: user-h...@cassandra.apache.org
