A major chunk of communication in an organization happens via email and any downtime in email can impact productivity and revenue. This is why availability and performance of Microsoft Exchange is vital for an organization that uses it.
To maintain uptime of Microsoft Exchange, it's essential that performance and availability is continuously monitored. But what are the parameters that have to be monitored? Here are some issues that can affect your exchange server performance and the parameters that need to be monitored to avoid those issues.
Storage Performance
As an IT admin, you would have seen cases where Microsoft Outlook users experience poor performance while trying to fetch emails from the exchange server. Storage Performance, i.e. input/output operations per second (IOPS), on the exchange server could be the culprit here. This is because IOPS defines how fast data - in our case email data - can be written to or read from the storage. Monitoring the performance of the Exchange storage lets you know of possible performance issues that can have an effect on mail fetching or sending.
RPC Threads
Do you often get support calls with users complaining that Outlook is unable to connect to their mailbox? One cause for this can be unavailability of Remote Procedure Call (RPC) threads. Outlook client connects using RPC threads to the Exchange server to perform operations, such as sending and receiving email, creating appointments, meetings and tasks, and so on. There's a limit in the number of available RPC threads on an Exchange server.
In cases where all RPC threads get used up, Outlook client automatically retries the connection until RPC threads are available making the user action slow. You can make RPC counters, such as RPC Requests, RPC operations/sec and RPC Averaged Latency counters throw alerts when the permitted limit is crossed by monitoring them. And once you receive an alert, you can restart the Exchange RPC Client Access service to free up RPC threads.
Something else is that an increase in the usage of RPC threads can also cause a bottleneck on the server’s resources, (RAM and CPU) thereby slowing down the server itself.
Replication
Data loss incidents, such as file corruption, water damage, human error, and so on can occur in an organization. All organizations should foresee these undesirable incidents and they need a database backup.
Most organizations run Exchange with the replication feature. This feature from Microsoft Exchange Server enables high availability for the Exchange Server's database. But just having this feature in your Exchange Server is not enough, ensuring the proper operation of replication is needed. An improper or fractional database backup is as bad as not having a backup at all. Thus copy status, copy queue, and replay queue for both the active and passive copies of all mailbox databases should be monitored to ensure there's no failure looming.
Further, replication status check is essential for factors like Active Manager, Cluster Service, and Replay Service, etc.
Storage Limits
When storage used to be expensive, Exchange admins used to limit the size of the mailbox storage. But now, because of cheaper storage space many admins decide not to have storage limits. This can cause the database size to grow and further on cause issues, such as a backup failure or increased restore time. Thus it is recommended to limit the mailbox size to minimize the time needed for data restore and reduce the probability of backup failure. Monitoring the mailbox size helps check if the applied rules for maintaining the mailbox size is operational.
Detailed monitoring and proper alerting for the above mentioned counters will help you take action before most undesirable events happen or get out of hand.
Praveen Manohar is a Head Geek at SolarWinds.
