This page says that the availability of an azure storage account is computed as (billable requests)/(total requests). Billable requests mean all the requests excluding those which experienced anonymous failures (except network errors), throttled requests, server timeout errors and unknown errors.
Now,what I see on the azure portal for my storage account is a straight line continuously at 100%, meaning that the account is available at 100% availability continuously. The line is without any break which means that the availability is being calculated continuously.
I know for sure that I don't throw requests to the storage account continuously. Then, how is this metric calculated for times when there are no requests?
Additionally, even a slight drop in storage availability means that some requests failed due to some server side issues. How can we ensure that these failed requests are retried and they pass?
When there isn't any incoming request, the availability is 100%. If your request encounters server side failures, you should retry the request via your code explicitly (in .NET client library, you can easily leverage RetryPolicy. See more details about RetryPolicy here).