Configuring a fail point on the config server can fail if the config server steps down while the command is being run. We should modify the fail point utils to use the assertRetriableCommandWorked helper to retry this command for these errors.

We should also add a new helper to the failpoint util to allow for setting a fail point for all nodes in a list, since this is a common pattern [1, 2] that also needs to rerun commands on retriable errors.