Add EarlyStoppingStrategy support for benchmarking

[esantorella]

Summary: This PR:

• Adds an early_stopping_strategy attribute to BenchmarkMethod. It is then passed to SchedulerOptions, and gets used by Scheduler (as in the benchmarking code never needs to reference or call the early_stopping_strategy)
• Changes seconds_between_polls on the ESS to 0, to match seconds_between_polls in SchedulerOptions. If this were not done, both would be set to the seconds_between_polls on the ESS, which by default is 300. Also, warn if this has been done.

Unrelated change:

• Correct a docstring in ThresholdEarlyStoppingStrategy to say it is checking for "worse" rather than "greater" values, since it does seem to account for whether the metric is to be minimized or maximized.

Differential Revision: D66010090

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D66010090

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 95.72%. Comparing base (7b671dc) to head (c519aab).

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #3071 +/- ## ========================================== + Coverage 95.69% 95.72% +0.02% ========================================== Files 488 488 Lines 49510 49558 +48 ========================================== + Hits 47381 47437 +56 + Misses 2129 2121 -8 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

---

🚨 Try these New Features:

• Flaky Tests Detection - Detect and resolve failed and flaky tests
• JS Bundle Analysis - Avoid shipping oversized bundles

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D66010090

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D66010090

[facebook-github-bot]

This pull request has been merged in facebook/Ax@e17f075cd940ff0827346d162e4e3d97fe9d179b.