Due to many factors, a TX may be initially reported as failed, but still be mined eventually. Such factors include gas prices, timeouts, etc.

Currently, our code deals with TX failures, but I'm not sure if it deals with TX that were initially failed, leading to a failure in the parent call, but end up succeeding in the background.
The following is a hypothetical example: Imagine that Alice uses grant(), creates a policy, sends the TX to PolicyManager, and after waiting for some time, we timeout and the call to grant() fails. It's not impossible, however, that the TX does succeed eventually. Is our grant() logic resilient to that? Would it be possible for Alice to recover the situation and resume the operation now that the TX exists? Or is she condemned to live with a zombie policy?

The case of Alice may be the most flagrant, but there are other places to where this can happen.

Related to #2070