API - Device Utilization

[RaphaelL]

Hi ,

I'm currently in a major incident involving 4 MX450 HUBs that are reaching 100% device utilization. 

We have monitoring in place , but the main issue is that statement : 

https://documentation.meraki.com/MX/Monitoring_and_Reporting/Device_Utilization

Note: The Performance API call returns latest performance score average from within the last one hour. e.g. If the Performance API is queried at 21:00, the API will return the performance score average from 20:00.

Does it really make any sense at all to have to stats from the last hour ?

Why can't I have the 'live' stats ?

We might have found the issue ( App deployement that is stopped ) but how to make sure ? I don't want to wait 1 hour...  

I know I can contact Support to tell me the advanced live stats ( PPS , BW , CPU usage and so on ) but it doesn't make any sense to me...

Anyone agree ?

[ww]

Agree. They should make a near realtime graph and i would like to see alerts if f.e. the avg of 5 min is above 90%

[BazMonkey]

Hi Raphael,
We've got a few large customers with MX450s that have been maxing out their 'device utilization' and when we contacted TAC they were able to provide detailed CPU reports for the MX Hubs showing the CPU is hitting over 100% regularly. This is causing latency issues, HA failover and some OSPF drops to its neighbouring core switch. We have optimised the MX's as much as we can and it did improve. But CPU has slowly crept up again. They are suggesting we fail over the to the standby MX as the first test. We'll do that tonight. The other customer we have they've offered an extra MX to load balance the traffic. Smells to me like there are known issues. 

Would be good to get an API call for it but I'm guessing they don't want us to know.

Let me know how you are going with a resolution. We'll being going vMXs in a few months so that might be our 'fix'
Cheers
Barry