Eliminate downtime during flush index operations

## Overview:
Running an index operation that flushes data is very disruptive. Since the whole index is deleted and then indexed, unindexed ES data is not available during runtime of the index. This can be quite problematic for sites with larger indexes as well as for sites that may heavily rely on Elastic search.

During flush index operations, if SearchPress could write to an auxiliary index and switch the indexed data after completion, this could allow for a little-to-no-downtime situation for sites.

## Proposed User story:
As a user initiating an index operation that flushes indexed data, I should see the Elastic search functionality experience little to no downtime.

## Proposed MVP:
- [ ] During index operations that flush data, an auxiliary index is written to.
- [ ] When index operations that flush data complete, the old index data is replaced with the auxiliary index data.

#### Nice to haves:
- [ ] An interface to trigger the index swap after a successful reindex.