You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A GameServerBuild became Unhealthy
A pod got evicted (because OOM), but the GameServer was not deleted because the GameServerBuild was Unhealthy and scaling was skipped
NodeAgent did not receive any heartbeats (since the Pod got evicted) so it marked the GameServer as Unhealthy
Increasing the CrashesToMarkUnhealthy did not incur deletion of the Unhealthy GameServer, since we're currently not doing that.
Scaling does not take Unhealthy into account, so eventually we ended up with an Unhealthy GameServer and less StandingBy than desired.
We should decide on what to do on these cases.
delete the Unhealthy GameServer
keep the Unhealthy GameServer alive for investigation
In scenarios that we have Unhealthy GameServers, GameServerBuild controller does not take them into account when scaling.
The text was updated successfully, but these errors were encountered: