Cosmos: Harden and merge tip-isa-Batch branch

[bartelink]

There's a tip-isa-batch branch, which generalizes the storage scheme to provide even greater efficiency in terms of read and write costs.

This work was shelved for prioritisation purposes in order to avoid having to make projectors deal cleanly with seeing an event multiple times.

Given a projector that does some practical de-duplication of events, its pretty feasible to re-introduce this storage optimization; the performance and RU cost reductions can be inspected by using Equinox.Tool run and/or dotnet new eqxtestbed against said branch.

As noted in #108, this storage generalization also provides benefits wrt competing writer scenarios.

[bartelink]

Removed milestone to reflect that, while this is unlikely to ever make it to a 2.x release (targeting Microsoft.Azure.DocumentDb V2), it's also not guaranteed to make a 3.x release as the Archival/Trimming facilities in #226 will likely yield a better overall solution. The other reason it's not a no-brainer to add back in is that the write amplification effects of having to read, write, and then have the CFPs read the redundant events are such that it requires a very specific combination of things (e.g. small total JSON length of events per stream) in order for this to be the best solution

[bartelink]

More reflections: Learnings and realizations from implementing Archival+Pruning actually make the case stronger for implementing this:

• The archiver has the opportunity and incentive to spend more time packing (events in tip and/or compressing event bodies) when writing to a secondary - the write amplification considerations don't apply to the same extent as they do on primary