Monitoring for Back Pressure thresholds - Exchange 2016

I had a situation recently where one of my Exchange servers started denying connections due to back pressure hitting a threshold. I noticed the Exchange MP didn’t catch this and I was wondering if anyone has come up with a clever way to catch this before it stops transport connections. I can of course have SCOM watch for the events but that’s more reactive. I’d prefer to get a notification when the disk space (for the threshold) is reaching it’s limits. In case anyone is unaware how this works, you could have an Exchange server with 50 gb free on the main binary volume but the size of the submitted messages waiting to be processed can reach it’s limit.



I suppose the first question is what action are you going to take if you see pressure rising on the server? If you are seeing it regularly it may be better to move the queue to a seperate drive with more space.

Some ways you could check are running a script to check and writing an event to the event log which you can pick up with Scom.


#Script to check for possible backpressure issues in Exchange 2016

#Import Exchange Module

Try {Add-PSSnapin Microsoft.Exchange.Management.PowerShell.SnapIn}
Catch {Write-error ‘MSExchange tools not found’}

[xml]$bp=Get-ExchangeDiagnosticInfo -Process EdgeTransport -Component ResourceThrottling
#Collect Diagnostic info for server
$issue = $bp.Diagnostics.Components.ResourceThrottling.ResourceTracker.ResourceMeter | Where-Object {$_.CurrentResourceUse -ne ‘low’}
#Check collected info for potential issues
if ($issue -ne $NULL) {Write-EventLog -LogName Application -Source MSEXCHANGE -EventId 4598 -EntryType Warning -Message ‘Check for backpresssure issues’ -Category 1 -RawData 10,20}

Thanks for responding. I’ll try this script. As for moving the database, I always can add more disk space since I’m virtualized but the issue is not the drive is running out as there is 50 gb free. It’s the backpressure calculation % of disk space is getting triggered. I’m looking really just to know when we are getting close so I can look further into the why here (whether it’s cpu/mem/disk latency etc). Since we are load balanced it’s odd to me that we are getting back pressure on one server but not the other.