Detect and report MicroBurst

triggers

#1

Has anyone work to create triggers to report on detecting MicroBurst and how high it goes along with how long it continues?

We have some EXA-1100 deployed in the Field when there are ‘issues’ that are hard to detect from the Central EXA device. One item that we are suspecting is some MicroBurst activity taking place and would to be able to provide better proof that it is happening and causing issues.

Have others studied this process? How have you detected it? How have you addressed it?


#2

The max bit rate for network throughput has a one-second resolution. You can see this on the Network dashboard within the System Dashboards folder.

I would also recommend navigating to Metrics => Applications => All Activity => Network & TCP to examine the trend for RTTs, RTOs, and stalls.

Hope this helps!


#3

I saw the Max bit rate - -thanks for letting me know it is at a one-second resolution.

Micro Burst is down to the 1 millisecond level - but I would be happy with the 10 millisecond level to let me know how often it is bursting for either a device or a sub-net


#4

FYI - - on some of our applications even at the 30 second interval level, we are seeing a 20::1 and 30::1 ratio of Peak :: Average levels.

No wonder we need to be concerned when we cross some low level utilization levels.