App Service Plans (Linux) restarts without planned maintenance

Question

Hi all,

All of our App Service Plans (Linux) in West Europe were restarted around June 3rd 10:00PM UTC. This caused a high CPU percentage and timeouts in our function apps.

As far as we know, there were no planned maintenances according to Service Health. We also didn't restart them ourselves.

Wondering if anyone has experienced something similar or if any experts have any suggestions on how to investigate these restarts?

Any help is greatly appreciated.

Thank you in advance!

Luna

Answer 1

@Luna Thanks for asking question! Could you please go to Diagnose and Solve Problems and look for a restart detector.

This tool can help you find out the possible causes of restarts.

For details refer to- Azure App Service diagnostics overview

Also, can you verify if you have enabled Auto-Heal rules? These rules will trigger a full restart, which will affect processes such as Application Initialization.

See- https://azure.github.io/AppService/2021/04/21/Announcing-Autoheal-for-Azure-App-Service-Linux.html

Please let us know.

Answer 2

@Luna Thanks for reply!

There is possibility that there is a high latency on the I/O to the storage. The high latency can cause a failover to read only storage, and then the container tries to write to the logfiles it fails. If this happens on startup or restart, then the container fails and Azure will restart it again in a few seconds.

Below are the steps you may use to monitor the I/O operations:

• From Azure Portal browse to your app, and open the “metrics” blade, you can aggregate the IO related metrics per app and setup an alert rule accordingly to be notified when the IO operations exceed a certain limit

• You can find the meaning of each metric on this link

Could you please confirm on app service plan. Scaling up might be helpful.

As then app service plan will have more resource after scaling up and this will lower the I/O latency and improve the problem with the restarts.