Exchange Server - Back Pressure

Overview

The Back Pressure monitoring test appears in the Exchange Functionality section for Exchange servers holding the Mailbox or Edge Transport role.

Monitoring Benefit

The Back Pressure test reports whether Exchange transport is currently throttling or rejecting mail because a monitored server resource is under pressure. Back pressure is Exchange protecting itself: rather than letting the server become unstable, transport slows or refuses new messages until the resource recovers. A red Back Pressure indicator therefore means mail is being rejected on that server right now, not that it might be at some point.

This matters because the symptom users report — mail queuing, delivery delays, senders receiving rejections — looks like a mail flow problem. The cause is usually a disk, memory or queue resource on the transport server, and back pressure is what connects the two.

How the status is determined

The ENow Client collects the data from each monitored server using PowerShell, reading the resource meters that Exchange transport maintains:

$bp = Get-ExchangeDiagnosticInfo -Server <ServerIdentity> -Process EdgeTransport -Component ResourceThrottling
$bp.Diagnostics.Components.ResourceThrottling.ResourceTracker.ResourceMeter

Each resource reports its current utilisation as Low, Medium or High. The indicator reflects the worst of them:

The test runs whenever the Exchange PowerShell agent jobs run, and at intervals of no more than 15 minutes.

There are no alert thresholds for Back Pressure monitoring. This is deliberate and is a common point of confusion — the Alert Thresholds section contains nothing for this test. The Low, Medium and High boundaries are Exchange's own, defined in EdgeTransport.exe.config on the server, and Microsoft advises against changing them rather than resolving the underlying resource problem.

In the ENow console, Back Pressure is listed with the alert category Exchange Server Functionality Back Pressure and the workload Functionality. The test is subject to Observation and Tuning, and it can be switched off for an environment from the Exchange section of Alert Intelligence.

Which resources are monitored

Exchange tracks these resources on Mailbox and Edge Transport servers. Any one of them going to Medium or High is enough to change the indicator, so identifying which one is the first troubleshooting step.

How do we verify the monitoring test results?

1. Read the resource meters on the affected server. Run the same call the ENow Client makes, from Exchange Management Shell on or against the server in question:

$bp = Get-ExchangeDiagnosticInfo -Server <ServerIdentity> -Process EdgeTransport -Component ResourceThrottling
$bp.Diagnostics.Components.ResourceThrottling.ResourceTracker.ResourceMeter

This shows each resource with its current utilisation and its pressure level, which tells you immediately which resource caused the indicator to change.

2. Check the Application event log. Exchange logs every pressure transition, so the event log tells you when the condition started and whether it has been flapping:

These are logged by MSExchangeTransport Resource Manager.

3. Establish what transport is actually doing to mail. The behaviour differs by pressure level, and knowing which applies sets the urgency:

• Medium — inbound connections are throttled and acknowledgements delayed, particularly from non-Exchange sources. Messages from non-Exchange servers and from the Pickup and Replay directories may be rejected, resubmission and shadow redundancy may pause, and message dehydration may begin.
• High — everything above, and transport also begins rejecting messages from other Exchange servers and from mailbox databases. Where version buckets are critically high, outgoing message processing and remote delivery may stop.

Common warning or error results, and potential solutions