Provide ability to customize number of retries per tag during repo mirroring

Issue

• When one needs to mirror a very large number of tags and some of them fail then the entire operation enters failed state. Since the mirror is in the failed state, even the successfully mirrored tags will be deleted/reverted to an earlier state. The user has to start the operation from scratch and it is very time-consuming.
• Do we have an option to avoid the above situation and retry the failed tags to success so the overall Repository Mirroring would be a success?
• We lost 90% of the mirrored data because a bunch of images failed to pull repeatedly. This is unacceptable from a user experience point of view. Can the number of retries be configurable, either via config.yaml or per repository?