As Exchange administrators we all dread the ticket that comes in that says “My Email is slow”. Exchange is such a complex system with so many moving parts internally as well as externally, which requires you to look at all possibilities with these sorts of tickets.
If an Exchange administrator suspects a performance issue (after ruling out network or client related issues), one of the first things to do is review the Exchange Performance counters. Exchange Performance counters can be reviewed using SCOM and for those not using SCOM, Performance monitor can be used as well. Using Exchange Performance counters can help determine if you’re having disks, CPU or even network latency issues. If you’re using SCOM to monitor Exchange performance, it’s pretty easy to pull up SCOM and run some reports however when you are not, it can be challenging to remember which counters to even start with.
What I’ve always found it to be a pain about these performance counters is that there is so many of them and which ones really matter for what I am trying to troubleshoot. I’m all about working smarter and not harder, so below I’ve listed some of the key performance counters you want to keep an eye on. Feel free to print this out as handy cheat sheet to refer too but if you’re looking for a more in-depth look at all the Exchange Performance counters visit Microsoft’s TechNet site .
Exchange Performance Counters:
§ Memory usage
· Memory\Available Mbytes (MB)
· Memory\Pool Nonpaged Bytes
· Memory\Pool Paged Bytes & Memory\Free System Page Table Entries
§ RPC Operations
· MSExchangeIS\RPC Averaged Latency – below 50ms
· MSExchangeDSAccess Domain Controllers\LDAP Read Time - below 50ms spikes no higher than 100ms
· MSExchangeDSAccess Domain Controllers\LDAP Search Time - below 50ms spikes no higher than 100ms
· Database disks
o PhysicalDisk\Average Disk sec/Read - less than 20ms
o PhysicalDisk\Average Disk sec/Write - less than 20ms
· Transaction Logs disk
o PhysicalDisk\Average Disk sec/Read - less than 5ms , spikes no higher than 50
o PhysicalDisk\Average Disk sec/Write - less than 10ms, spike no higher than 50ms
· Log buffer
o Database\Log Record Stalls/sec – below 10 per sec no higher than 100 per sec
· Processor\% Processor Time (_Total) , ideally should be less than 90%