Hi Scott, I think I'm looking at a more comprehensive look at performance. Both for general monitoring (like you mentioned) as well as specific performance monitoring that will help determine the cause of performance bottlenecks and spikes.
Just a few comments for examples... > >1) What are the largest / larger known implementations of IMail? > > There are quite a few that process 100,000+ E-mails/day, and some with as > many as 100,000+ users. I don't know what configurations they are running, > though. This is similar to what Ipswitch says. The problem is that 100,000+ is pretty open ended, so I can only assume that it will handle up to 100k. I'd love to find out what those limits are and I want to help provide some docs for people who are considering IMail for large implementations. Even for smaller implementations, it would be nice for people to know how to size their system so they're not wasting cash on Xeon boxes when they could get by with a PIII. I'm currently consolidating nearly 140 autonomous IMail servers and Ipswitch had little documentation at all to help guide us with hardware requirements. Without prior knowledge, we've had to discover how well the product scales as we go (support hasn't been very useful either). > >2) What is the best way to conceptually measure the performance of > >various components of an IMail Server? Specifically what metrics > >determine the actual performance of a system and what factors should be > >considered for determining performance bottlenecks and threshholds > >(processor / network / memory / io utilization, user response time, SMTP > >delivery time, messages processed per minute / day, external user > >databases, etc.) > > The key for mail transfer and scanning is the number of E-mails > sent/received per day. I hope to document a full range of performance metrics, so as to include hardware performance as it pertains to mail sent/received per day as well as per minute to account for peak usage performance constraints. > >b) What type of information would you like to see on a best-practices > >performance guide? > > "If CPU usage is high, use Task Manager, go to the "Processes" tab, and > click on the "CPU" (not "CPU Time") button to sort the processes by the > percentage of CPU time used. The processes towards the top of the list are > using the most CPU time". > > Although this is generic to any CPU usage issues, it's amazing how often > people know that their CPU is at 100% but don't know to do the above. I couldn't agree more, it amazes me how often the basics are overlooked. Although I would also like to provide some intermediate how-to docs to help people troubleshoot specific performance issues. This might include how to determine for certain if it's a storage issue, network bandwidth issue, application issue, or faulty / abusive mail client issue. Even more detialed how-to's such as setting alerts in the performance monitor to trigger thread specific monitoring to determine which user is causing the POP3D service to spike abnormally. I would also love to find a stress tool to push IMail's limits. If anybody knows of a free tool, please let me know. Thanks, Ives To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/ Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/
