Vector based on lgalloc for flat container

[antiguru]

Introduce containers that allow flat container regions to spill to disk.

Depends on:

• Flat container storage abstraction TimelyDataflow/timely-dataflow#574
• Generalize flat container implementations to less specific regions TimelyDataflow/differential-dataflow#518
• Separate data storage antiguru/flatcontainer#19

Motivation

Tips for reviewer

Checklist

• This PR has adequate test coverage / QA involvement has been duly considered. (trigger-ci for additional test/nightly runs)
• This PR has an associated up-to-date design doc, is a design doc (template), or is sufficiently small to not require a design.
• If this PR evolves an existing $T ⇔ Proto$T mapping (possibly in a backwards-incompatible way), then it is tagged with a T-proto label.
• If this PR will require changes to cloud orchestration or tests, there is a companion cloud PR to account for those changes that is tagged with the release-blocker label (example).
• This PR includes the following user-facing behavior changes:

[antiguru]

This PR introduces a small performance regression, which can be explained (in the future prevented.) At the moment, we're using vectors to store on dataflow edges, which have the nice property that they can be preallocated to a specific capacity. This concept doesn't (yet) exist in flat containers. Ideally, we'd like to preallocate each region to its desired size, but without knowing the actual data, we cannot know the size of the regions.

The performance regression only affects the introspection dataflows, and seems to be in the range of 1-2%.

A potential way forward is to introduce a TryPush (c.f. antiguru/flatcontainer#54) trait that guarantees pushing without reallocation, combined with a way to preallocate regions with sufficient capacity, but more work is needed to enable this.