Hello Our environment is Exchange 2007 SP1. I had some questions about RPC Latency which I can't seem to understand the basis of. 1. What exactly is RPC latency? Could you say it's the 'speed' of a connection from an Outlook client to Exchange mailbox. 2. Is there a way to measure RPC latency? 3. What should be the threshold? 4. What can affect RPC latency?

RPC latency ( Remote procedure call latency , outlook uses rpc protocol to connect to exchange ). The more of this can affect your clients accesssing Exchange server You can set Performance counters to measure RPC latency., ( tools option in emc should help) Many factors other than network connectivity, not so good database health to Av scanning DB will affect RPC latency. You can also increase diagnostic logging for rpc counters to check individual sessions.

On Fri, 15 Oct 2010 18:23:20 +0000, Sheen1990 wrote: >Our environment is Exchange 2007 SP1. I had some questions about RPC Latency which I can't seem to understand the basis of. > >1. What exactly is RPC latency? Could you say it's the 'speed' of a connection from an Outlook client to Exchange mailbox. No, it's the time it takes to 'process' a RPC packet. >2. Is there a way to measure RPC latency? Use Performance Monitor. "MSExchangeIS \ RPC average latency" is the counter. >3. What should be the threshold? Not more than 50ms. >4. What can affect RPC latency? The ability of the server to handle the workload. That might be IOPS (too many for the disks to handle), but it may also be CPU (the ability to multi-thread), or memory (more is good because it reduces the need for I/O -- refer to IOPS). --- Rich Matheisen MCSE+I, Exchange MVP --- Rich Matheisen MCSE+I, Exchange MVP

Thanks Rich. One last question, if we were experiencing an issue and wanted to check RPC latency, would we need to increase diagnostic logging, or can we just run Perfmon straight off with this counter? And - if the value was more than 50ms, does this mean that the problem is Exchange server related (i.e not networking, client-side issues etc)? Thanks again!

On Sat, 16 Oct 2010 19:00:16 +0000, Sheen1990 wrote: >One last question, if we were experiencing an issue and wanted to check RPC latency, would we need to increase diagnostic logging, or can we just run Perfmon straight off with this counter? I'd go get a copy of Performance Analyzer of Logs (http://pal.codeplex.com) and install the prerquisites and then the tool. Then get the Exchange 2007 PerfWiz and us it to set up the perfmon on the server: http://blogs.technet.com/b/mikelag/archive/2008/05/02/perfwiz-replacement-for-exchange-2007.aspx Now, let it run for a while. It's a good idea to do this before you have a problem so you know, more or less, what things normally look like. Run the logs through PAL and address the stuff in red first. :-) Oh, you should run ExBPA too. There's no sense trying to find wht's out of spec on your own when someone's already done it for you. >And - if the value was more than 50ms, does this mean that the problem is Exchange server related (i.e not networking, client-side issues etc)? If that's where you're measuring, yes. --- Rich Matheisen MCSE+I, Exchange MVP --- Rich Matheisen MCSE+I, Exchange MVP