I am running boundary 0.7.4 and it is very slow, almost to the point of being unusable. What should I be looking at for scaling? My controller is sitting at a pretty average load of 0.3, and my worker is sitting at a load of 0.1. This feels like its not CPU bound, and both are on a 10G network, so I really don’t think its network bound either. I don’t think this is a memory issue but I unfortunately don’t have long term data for that.
Do you have any interesting events in the output from the controller or worker? We’re fixing some database issues in a release that should be later this week that have caused some deadlocks in the DB which could have this end result.
Overall I didn’t see any particularly interesting things in the log. I think the most interesting thing I’ve discovered recently is periodically controllers will just crash with no log output. I have put more restart logic in place and it appears to work more reliably after doing that.
It mostly works. I think there was some stability issue with 0.7.4 as well, now that I have 0.7.6 running with runit and it appears to mostly just work now.