Azure App Service is a service used to create and deploy scalable, mission-critical web apps.
Thank you for posting your question on Microsoft Q&A.
It appears that the App Service Plan may be hosting more active sites than recommended. This can lead to higher overall CPU usage compared to the sum of individual application CPU usage due to kernel services are also included in the overall CPU usage. These can be higher when system processes call APIs frequently, such as I/O, networking, and thread context switching.
Possible steps to consider:
Scale Up the App Service Instance Scaling up the App Service instance increases the available CPU, memory, and disk space to handle additional workload.
Split Applications into Separate App Service Plans To avoid one application impacting the performance of others on the same plan, consider moving production apps into a separate App Service Plan. This ensures that each app receives sufficient dedicated resources.
As a general guideline, here are the recommended maximum active sites per worker size:
| Worker Size | Max Sites |
|---|---|
| Small | 8 |
| Medium | 16 |
| Large | 32 |
If Application Insights Profiler is enabled, you may also try disabling it and observe the impact.
From the High CPU Analysis, high memory usage has also been detected on multiple instances. This usage is not limited to a single app or process, but rather multiple apps and system processes together (for example: mi-ministry, mi-tasks, mi-admin-api).
In such scenarios, scaling up the App Service instance or splitting the applications into separate App Service Plans can help improve stability and performance.
Thanks
Vimal Lalani