03-09-2017 09:01 AM
In our env, we have bottleneck alert coming as below .
Latency bottleneck on E-Port xx/yy. 0.33 pct. of 300 secs. affected. Avg. delay 2754 us. Avg. slowdown 644.
Please can someone explain what do the below fields indicate and how they are calculated ??
03-10-2017 12:28 AM
From the FOS 7.3 Message Reference manual at http://www.brocade.com/content/html/en/message-reference-guides/fos-741-messageref/wwhelp/wwhimpl/js/html/wwhelp.htm#href=FOS_messages.8.003.html
"Latency bottleneck on port <slot number>/<port number within slot number>.
<percentage of seconds affected by latency bottlenecking> pct. of <observation
period over which the percentage of affected seconds is reported> secs. affected.
Avg. delay <observed average time between frames during affected seconds> us. Avg.
slowdown <observed throughput drop factor during affected seconds>."
So the Average delay is the slow down in average between frames sent out during the affected 33% of 300 seconds or 100 seconds. The average slow down under these 100 seconds affected, is a factor of 644 times (slowdown). The value are calculated based on the buffer credit to zero values, which is measured every one second interval. So, for each second, we look at the amount of buffer credit to zero - notice that an increment of one means a 2.5 us delay, which mean for average 2754, we have in average of 1101 buffer credit to zero which incremented during the 100 seconds impacted during the 300 second intervall, or in total 110100 increment during these 100 seconds.
Some furthe details can be found in the FOS 7.3 Admin Guide at http://www.brocade.com/content/html/en/administration-guide/fos-730-adminguide/GUID-D41B5965-6282-417B-B8FC-9B8492578676.html.