When index validation finds an inconsistency it has to make a second pass on the index/collection to identify which records are missing/extra. This part of validation is memory intensive and it's hard to determine how much work will be done by looking at the logs.

LimitMemoryForSecondPhase is where we try to estimate the memory usage and limit it if needed. It would be useful for us to log the number of buckets being examined and the cumulative size of those buckets