|
IDEAS ONLY - I'm not a perf!
There are the canned dozen performance monitor configs, and then there's
the EPERFMON.HLP file that had some of these descriptions. I guessed
on what they could mean - my comments in ().
Ralph, where are you?
MTA
WorkQueueLength - outstanding messages to process
(consistently high numbers may indicate insufficient
processor capacity or indirect bottleneck like sluggish
store access on local system)
MTA
MSExchangeMTA Connections object:
Associations - how many open connections to other entities
(excess numbers could indicate need to move users onto
same server that are generating most traffic?)
Queue Length - number of items queued to other entities
(if you can't communicate fast enough to other servers
what is the bottleneck?)
IS
MSExchangeDB
Buffer I/O Queue Length - Outstanding I/Os to process
(need some more RAM, or adjust Registry entries for
increased buffers beyond standard install or perfopt
numbers)
Average Time for Delivery - avg time last 10 msgs took to get from
is to MTA
Average Time for Local Delivery - avg time last 10 msgs took to
be registered in IS and delivered to all recipients
(If these are on the increase, might suggest need for faster
disk subsystem and more CPU)
Send Queue Size
Receive Queue Size - establish baseline, then monitor for steady
increase indicating need for hardware upgrade or move
of some users and services to a subsequent server.
|