Rollback in the event of convergence failures

[rmc3]

It should be possible to roll back to a known-good configuration in the event of an update while converging changes to an AWS organization. For example, the tool should capture a model of the AWS Organization immediately before beginning convergence. It should test that it is able to converge to that state both from the configured state the target converged state before beginning convergence. In the event that an error is encountered, it should attempt to converge back to that captured model.

[rmc3]

This feature suggestion needs major reconsideration before being pursued.