Run full-history benchmark commands in CI to catch perf regressions #775

TL;DR

Run the full-history benchmark commands (#771) automatically in GitHub CI and compare results across versions, so performance regressions are caught before they land.

Scope

• A CI job that runs the benchmark commands on a defined dataset.
• Compare results against a baseline and flag regressions.

Depends on

• Full-history benchmark commands — ingest across all data types #771 (the benchmark commands).

(Full-history analog of #708, which does the same for the v1 Go benchmarks.)