hive-database unhealthy in consul
Closed this issue · 2 comments
pdmthorsrud commented
Hive-database is unhealthy at random intervals in consul. Might because pg_isready has a timeout of 2s, which might be too short because the system is fairly slow atm with everything provisioned.
fredrikhgrelland commented
It is 3 seconds. https://www.postgresql.org/docs/12/app-pg-isready.html Bet we should increase it to 5?
Interesting. I was only talking about the titmeout from nomad:
After changing that one to a higher number it seems to be stable on my end. I guess that means 3s is actually enough. Change it to 5s to be on the safe side? There seems to be a lot of extra delay because the system will run slow when a full stack is provisioned due to the limits of our VDIs.
fredrikhgrelland commented
Yes - increase timeout to at least 5 sec.