Retry container creation on failure

We should add some internal retry logic when creating a tenanted container (for any of the storage types) fails. This will allow clients to not require their own retry logic for foreseeable transient issues.

No files, tests, or entry points are named. Start by locating the tenanted-container creation paths for each storage type, then determine the intended retry limit and which failures are transient; done means all supported storage types retry creation consistently without requiring client-side retries.

Written by the indexing model from the issue text.