Currently there's no retry mechanism built into the mongo-container pipeline's push step. This results in failures due to network flakiness. Add in a retry option since there isn't likely to be any other way to improve this.

See BF-30221 for an example of this