In order to support ephemeral infrastructures a filter was added to exclude stopping and stopped VMs
The filter needs at least 25 minutes for vm state to be available, so a maxDelay of 15m was added on top of the 10m of lasting
Problem is this maxDelay is confusing and might not be enough if data is really delayed, so moving to a lasting of 25m (and reverting maxDelay to auto) should help a lot.
Furthermore a variable heartbeat_exclude_not_running_vm is added in order to switch off the filtering if not needed.
In order to support ephemeral infrastructures a filter was added to exclude stopping and stopped VMs The filter needs at least 25 minutes for vm state to be available, so a
maxDelayof15mwas added on top of the10moflastingProblem is this
maxDelayis confusing and might not be enough if data is really delayed, so moving to a lasting of 25m (and revertingmaxDelaytoauto) should help a lot. Furthermore a variableheartbeat_exclude_not_running_vmis added in order to switch off the filtering if not needed.