We've recently turned off bottleneckmon and turned on MAPS FPI monitoring. We're used to the idea of latency being when the buffers fill up and frames are lost (and so IO_FRAME_LOSS messages make sense), but we're not entirely sure what IO_PERF_IMPACT really means. We get a lot of alerts for one particular port that read similar to:
2016/06/28-07:27:13, [MAPS-1003], 10650, FID 128, WARNING, c80fsw113, F-Port 43, Current Value:[DEV_LATENCY_IMPACT,IO_PERF_IMPACT, 31.2% of 10 secs], RuleName=LHR_Latency_Perf_Impact.
But this doesn't cause ANY issues - there are no errors in porterrshow, portstatsshow, portshow, no impact to the connected server.
Is this normal? Is this something we can just ignore? What does the 31.2% of 10 seconds actually mean? Might this be something to do with TX Credit Zero?
Any advise appreciated - the manuals and Google don't shed much light.