This one error probably needs its own section in docs, there are a lot of potential causes. The key thing to fixing it is to understand how Umbraco uses the NuCache file.
When Umbraco boots it opens the NuCache file, and locks it open, until it's shut down.
On Azure, a new instance might get spun up before the other shuts down (see WEBSITE_DISABLE_OVERLAPPED_RECYCLING,
https://learn.microsoft.com/en-us/azure/app-service/reference-app-settings?tabs=kudu%2Cdotnet#logging) - this can happen whether you're scaling or not, and should be disabled for Umbraco sites in Azure.
Also on Azure, the default filesystem is actually a remote file share which can be flaky - sometimes files are not released when they're finished with. This is much (much much) more likely when Umbraco fails to shut down properly (which is usually due to scheduled/background tasks that developers have failed to add graceful shutdown logic to). For that reason the docs recommend that you use these settings:
https://docs.umbraco.com/umbraco-cms/fundamentals/setup/server-setup/azure-web-apps#recommended-configuration this stores the cache and Lucene on the fast local SSD which doesn't have the same problems.