Provide means of identifying when CloudMapper runs out of memory and document this possibility

As seen in the comments of #666, CloudMapper can run out of memory. Have a way to identify when this has happened. What signs can someone look for and can I detect this is going to happen before it does and warn? Also maybe give some guidance, especially for the nightly auditor, about when this situation should be expected and how to avoid it.

Review the comments on #666 and the nightly auditor context first. Determine which observable signs indicate that CloudMapper has run out of memory, whether an early warning is feasible, and what guidance should be documented for expected conditions and avoidance. Done means the behavior is identifiable and the relevant usage guidance is documented.

Written by the indexing model from the issue text.