GCS: Retrying failed requests

[defined-user]

Working with a GoogleCloudStorage API may fail due to some transient server side errors, such as TooManyRequests, ServiceUnavailable, InternalServerError, etc. Retrying failed requests is recommended by Google for some specific errors. As an example, gsutil tool, which is written in Python, has it's own retry handling strategy (more information on what errors are considered to be retryable can be found there). These errors could be handled with an exponential backoff retrying algorithm. In case, if that does not help, it would still prevent the client from making too many requests and and retrying indefinitely.
Beside that, aforementioned gsutil tool also retries on transient network errors, and I consider the one specified in @and3rson's issue #735 to be one of those. Using a standard Google's retry handler allows to handle such exceptions by appending the list of exceptions that must be handled.

[defined-user]

Closing it, as there has been no feedback from the maintainers on this issue and related PR for almost a year.