| < Day Day Up > |
|
The Exchange administrator should monitor trends to detect changes in the use and performance of the Exchange messaging environment over the course of time. The objectives of trend monitoring are:
Early detection of messaging service degradation
Analysis of stored performance information to identify, analyze, and react to changes
Collection and correlation of historical data to assist with detection, analysis, and correction of systemic faults
Comparison of actual service levels with the targets established in SLAs
Data collection and analysis to assist with capacity planning
Ability to forecast future growth to determine hardware and software needs
Monitoring to identify trends requires that data be collected at regularly scheduled intervals to provide the source data for producing reports. The scheduled data collection also provides the database for ad hoc queries, report generation, and analysis.
The administrator must carefully balance the quantity of collected data against the impact that the data collection will have on network and server performance. In general, it is best that the combination of health and trend monitoring should add no more than a 3% to 5% performance penalty during normal operation. Unlike health monitoring where a backup monitoring site is recommended, a backup monitoring site is not recommended for trend monitoring, because it will cause an unnecessary load on the network and the servers.
The Exchange administrators should establish policies regarding the type and quantity of data that should be collected for trend monitoring. Tables 11.2 through 11.4 describe some recommended counters for Exchange messaging data collection. These should be periodically reviewed and adjusted to ensure that they are providing the necessary information needed for reporting and analysis, without placing an unnecessary adverse strain on the network and Exchange servers.
Object | Counter |
---|---|
Memory | % Committed Bytes In Use |
Memory | Pages/sec |
Paging File | % Usage |
Processor | % Processor Time |
SMTP Server | Message Bytes Sent/sec |
Object | Counter |
---|---|
LogicalDisk | % Free Space |
LogicalDisk | Free Megabytes |
MSExchangeIS Mailbox | Total Size of Recoverable Items |
MSExchangeIS Public | Total Size of Recoverable Items |
Object | Counter |
---|---|
MSExchangeIS | Active User Count |
MSExchangeIS | RPC Requests |
MSExchangeIS | User Count |
MSExchangeIS Mailbox | Messages Delivered/min |
MSExchangeIS Mailbox | Messages Sent/min |
MSExchangeIS Public | Messages Delivered/min |
MSExchangeIS Public | Messages Sent/min |
MSExchangeIS MTA | LAN Receive Bytes/sec |
MSExchangeIS MTA | LAN Transmit Bytes/sec |
MSExchangeIS MTA | Message /Sec |
MSExchangeIS MTA | Message Bytes/Sec |
SMTP Server | Message Bytes Received Total |
SMTP Server | Message Bytes Received/sec |
SMTP Server | Message Bytes Sent Total |
SMTP Server | Message Bytes Sent/sec |
LAN, local area network; RPC, Remote Procedure Call; SMTP, Simple Mail Transfer Protocol |
The Exchange administrator should regularly analyze the usage of the Exchange messaging environment. This can be done most easily by periodically logging and analyzing the message traffic on a sample set of Exchange servers. These logged messages should be examined to determine the average size of messages, the typical number of recipients, the location of the recipients (e.g., internal Exchange users, internal users on other mail systems, external Internet users), the type and size of attachments, and the typical message attributes. This information can then be used to improve the accuracy of capacity planning. It is also often useful to let the funding business units understand how the messaging system is being used.
| < Day Day Up > |
|