1 Reply Latest reply on Jan 13, 2012 7:05 AM by RayCologon

    Performance Counters - Benchmark

    fm-adm

      Hi,

       

      I came across this article which discribes which counters to monitor on FM server:

      http://help.filemaker.com/app/answers/detail/a_id/9648/~/filemaker-server-for-windows-performance-monitoring

       

      As helpful as this was, I was not able to find what values are considrered good/bad for each counter, does anyone know where I can find this info?

       

      Thanks!

      AK

        • 1. Re: Performance Counters - Benchmark
          RayCologon

          fm-adm wrote:

          Hi,

           

          I came across this article which discribes which counters to monitor on FM server:

          http://help.filemaker.com/app/answers/detail/a_id/9648/~/filemaker-server-for-windows-performance-monitoring

           

          As helpful as this was, I was not able to find what values are considrered good/bad for each counter, does anyone know where I can find this info?

           

          Thanks!

          AK

           

          Hi AK,

           

          I don't think that benchmarks of the kind you're asking for exist, or (if they did) would be of much real value. 'Good' and 'Bad' are relative terms. The issue is one of acceptable performance, where acceptability may vary widely according to the processing loads called on for the databases being hosted, the numbers of users and the expected system responsiveness for different types of transactions and procedures.

           

          Not infrequently, the interface and workflow of a solution also determines (or at least affects) perceived adequacy or otherwise of performance, so even for similar types of solutions hosted in equivalent environments, adequate (and therefore 'good') performance in one case might be considered unacceptable ('bad') in another. Moreover typical network response times tend to affect user expectations in a similar way, so if users are accustomed to network bottlenecks, they will be less aware of (and more tolerant of) performance issues around database i/o or load.

           

          It's also not helpful to benchmark systems against "what's possible", since configurations vary widely with a plethora of variables all affecting performance. There are some vendors who quote performance statistics in their sales materials, but I doubt that will give you a meaningful basis for comparison.

           

          What you *can* consider doing is aggregating performance data from your own systems and solutions over time so that you can arrive at meaningful comparisons and measures, with points of reference that are pertinent to your environment, requirements and user expectations. That's how counters become useful - you can compare the response times and event counts for period with those from other similar periods (for the same setup) to determine where problems lie and what to do about it.

           

          Regards,

          Ray

          ------------------------------------------------

          R J Cologon, Ph.D.

          FileMaker Certified Developer

          Author, FileMaker Pro 10 Bible

          NightWing Enterprises, Melbourne, Australia

          http://www.nightwingenterprises.com

          ------------------------------------------------