Azure App Service Plan CPU Spikes to 100%

Tahami Rizwan 20

I've been experiencing intermittent CPU spikes to 100% in my Azure App Service Plan, but when I investigate the individual App Services within that plan during the same time range, the CPU usage appears to be normal. I've also closely examined function calls and SQL queries using both Application Insights and Stackify, and they all seem to be operating normally during these spikes.

I'm struggling to identify the root cause of these CPU spikes. Could anyone suggest what might be causing this issue or how I can further investigate it? Any help would be greatly appreciated!

Thanks

Lex Li (Microsoft) 4,742 Reputation points Microsoft Employee

2023-09-09T16:54:58.61+00:00

Which App Service plan are you using? There are guides like https://azure.github.io/AppService/2020/10/05/Diagnose-App-with-High-CPU.html that might get your started.
Tahami Rizwan 20 Reputation points

2023-09-11T04:12:14.0166667+00:00

I am using P1V3 app service plan and I have check app service CPU it is normal
Lex Li (Microsoft) 4,742 Reputation points Microsoft Employee

2023-09-11T05:26:02.6666667+00:00

Then can you describe more about "intermittent CPU spikes to 100% in my Azure App Service Plan"? Where exactly did you get that notice?
Ben Gimblett 3,410 Reputation points Microsoft Employee

2023-09-11T09:54:03.03+00:00

To add to @Anonymous comment and making an assumption for just now that the source is Azure Metrics --- the CPU metric for the App plan is "CPU Time" the metric for a web app is "CPU percentage"

Time and percentage are different measures

More importantly, the CPU Percentage metric (for the web app) is an aggregate of the instances running against the web app plan (e.g. 2 instances would equal 2 underlying worker VM each identically running a copy of the actual app or apps and load balanced). To get a clearer picture of potential spikes you need to
(a) "split" the CPU percentage metric so you see the slice for each instance (can do this easily via the metric UI in the portal)
(b) look at short time windows relative to the metric capture frequency (zoom in) and compare metrics in the same small time window , because the longer the time the less easy it is to see anomalies which otherwise get averaged out (in general Az monitor collects metrics on a 1min frequency)
Tahami Rizwan 20 Reputation points

2023-09-12T04:19:04.6633333+00:00

In app service plan I am facing CPU spike as image is attached below but when I go CPU usage in app service it's utilization is at normal level
SnehaAgrawal-MSFT 18,366 Reputation points

2023-09-15T06:06:44.1333333+00:00

@Tahami Rizwan If you are still facing issue, could you please reach out to us by sending an email with subject line “Attn:Sneha |thread title” to AzCommunity[at]microsoft[dot]com referencing this thread, to investigate deeper on this.
galletto 186 Reputation points

2024-02-26T10:03:54.74+00:00

@SnehaAgrawal-MSFT @Tahami Rizwan , I'm having the same issue: an app service plan is showing intermittent peaks of 100% CPU usage. I can't relate it to any website activity or webjob. I's happening in the middle of the night, as well as during daytime here. Did you find out anything about this? Thanks, Will de Haan
Saad Awan 10 Reputation points

2024-04-26T13:08:19.3766667+00:00

@SnehaAgrawal-MSFT , Yes, I am also experiencing the same issue with my 4 different Production App plans hosted in the AUS EAST region. After 12 to 9 hours or randomly, the CPU spike touches 100%, either during production or non-production hours. When I check the logs at that specific time in the ApplicationInsight, there are no errors, exceptions, or performance issues. My management is asking for the reason for this spike.
Saad Awan 10 Reputation points

2024-04-26T13:09:27.98+00:00

Yes, I am also experiencing the same issue with my 4 different App plans hosted in the AUS EAST region. After 12 or 9 hours randomly, the CPU spike touches 100%, either during production or non-production time. When I check the logs at that specific time in the ApplicationInsight, there are no errors, exceptions, or performance issues. My management is asking for the reason for this spike.

1 answer

Muhammad Ali Khan 0 Reputation points

2024-04-22T14:25:17.81+00:00

I'm working for an IT company, facing the same issue with an Azure service plan (P1V3) shared by 12 APIs hosted on Azure app services. 100% Spikes are quite random. When the request responses are examined no major issue is found during the spike time, Now how do I explain and justify to the management, who are not too technical people and they trust the Azure Metrics graph more than anything
Please sign in to rate this answer.

0 comments No comments
Sign in to comment