Email is one of the most heavily used communications method which makes your Exchange servers critical to your business.
The health and performance of your Exchange servers should be a top priority, and this means that you must monitor the server performance as part of your routine so that problems can be discovered early and resolved before they begin to make a serious impact.
Here are 5 performance counters to monitor on your Exchange servers today.
1. % Processor Time
This counter shows the amount of time that the CPU is processing a task. This counter should typically be below 75%, although it may run higher during heavy workloads such as backups. If the processor time is consistently high you will want to look into which processes are utilizing the CPU the most.
2. Processor Queue Length
When instructions are sent to the CPU they go into a queue to be scheduled for execution. This counter shows the length of that queue, and should ideally be no higher than 5 for each processor in the server.
When this counter is above the ideal threshold along with a high % Processor Time it indicates that the server workload is too high for the CPU resources available.
3. Memory Available MBytes
This counter shows the amount of memory that is not in use and is available for new tasks or processes, and should be at least 100Mb at all times.
If it drops below that threshold then the server memory is inadequate for the workload, and excessive disk utilization may result due to heavy pagefile use to compensate for the memory shortfall.
4. Memory Pool Paged Bytes
There is no threshold to watch for on this counter; however, it should be monitored for any increases over time which indicates a memory leak may be occurring.
5. Physical Disk Average Disk Queue
This counter should be monitored for each individual volume on the server, not for the total amount. In general a disk queue of 2 or less is acceptable. A higher disk queue, especially if it is sustained for a lengthy period, indicates that disk I/O is exceeding the performance capabilities of the disks themselves.
This can be normal during heavy disk operations such as backups, but should always be investigated to verify the cause. When a high disk queue is noticed the first step is to break down whether it is being caused by read or write operations by inspecting the Disk Read Queue and Disk Write Queue.
A Basic Performance Snapshot
These 5 performance counters give you a very basic snapshot of your Exchange server performance. This is very useful even when no one is reporting any problems with the server, because it is important to have performance benchmarks during “good” times to compare to the times when a problem is reported.
If any of the counters indicate a problem in that area then you can start more comprehensive monitoring of related counters to narrow down the root cause of the issue so that it can be resolved.