persist: truncate String and Vec<u8> stats to a maximum byte length

[danhhz]

These stats are inclusive lower and upper bounds, so we can (carefully) truncate almost any String or Vec value to produce one that's smaller but still a relatively tight bound.

There is TONS of stats pruning left to do (including given the ProtoDatum and Json stats the same treatment), but this is enough to get CI (which sometimes inserts 100MiB strings) to pass with the stats collection ff on, so it seems like a nice start.

Touches #12684

Motivation

• This PR adds a known-desirable feature.

Tips for reviewer

Checklist

• This PR has adequate test coverage / QA involvement has been duly considered.
• This PR has an associated up-to-date design doc, is a design doc (template), or is sufficiently small to not require a design.
• This PR evolves an existing $T ⇔ Proto$T mapping (possibly in a backwards-incompatible way) and therefore is tagged with a T-proto label.
• If this PR will require changes to cloud orchestration, 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:

[bkirwi]

LGTM!

(I think it would probably be okay to just not write stats for columns like this. But this is of course even better!)

[danhhz]

yeah, I think we want a general mechanism for pruning stats for a column, but it's not clear to me yet exactly what the heuristics for that should be (I'm tempted to say that's probably something to be vetted during the design doc process, which it seems about time to circle back to), whereas this felt like a straightforward win. TFTR!

[pH14]

The logic to get an upper bound from a UTF8 string actually reads much simpler than I expected, very nice!

Wouldn't mind a proptest that generates strings/Vec and verifies the lower and upper bounds work as intended over a myriad of inputs

[danhhz]

The logic to get an upper bound from a UTF8 string actually reads much simpler than I expected, very nice!

lol that took me 7 hours to write and I almost gave up at one point XD

[danhhz]

Wouldn't mind a proptest that generates strings/Vec and verifies the lower and upper bounds work as intended over a myriad of inputs

done! #18805