[Backport MB-56365 to 7.2.1] Enable index bloom filter by default

Index bloom filter helps in the following cases:

1. Partial indexes - If there are many partial indexes on the bucket, indexer will have to do a back-index look-up even if the mutation does not belong to the index (as the type field of the document might have changed - and indexer has to delete the old value in case it changes). These back-index lookups require fetching page from disk. The random IO done for back-index lookup can become a bottleneck. CBSE-14023

2. Insert workload - For the insert only workload or a dominant insert workload, indexer will need to do a back-index lookup as it cannot differentiate between an insert vs. update coming via DCP. This slows down the incremental mutation processing.

It is better to enable the bloom filter by default. With in-memory compression for plasma in 7.1, the memory requirements have been reduced from 20% RR to 10% RR and this is a reasonable trade-off to allow extra memory for bloom filter given its benefits.

This can be considered for both on-prem as well as Capella.