Implement re-try strategy / pause - resume on re-index

[rayrutjes]

On the backend, when waiting for re-indexing to be finished, it might happen that one of the queries to the backend fails.

We have recently had a user victim of a gateway timeout.

In that case we should probably want to re-try the last query at least 2 times at couple seconds interval.

[rayrutjes]

This would also solve the when a given post is problematic during indexing. It would skip the whole page, but that would be already better than skipping till the end of the indexing.

[ethansmith]

This case (via a query error) has happened to me a few times, and has usually gone unnoticed (until later) with large record sets. In no case has an error notice been generated.

Some alternative behavior, or even just an error dialog, would be great!