[WIP] indexheader bucket reader scratch reader benchmark and stats

[francoposa]

What this PR does

Which issue(s) this PR fixes or relates to

Fixes #

Checklist

• Tests updated.
• Documentation added.
• CHANGELOG.md updated - the order of entries should be [CHANGE], [FEATURE], [ENHANCEMENT], [BUGFIX]. If changelog entry is not needed, please add the changelog-not-needed label to the PR.
• about-versioning.md updated with experimental features.

[francoposa]

Right now, collecting the caching bucket stats in this way is not interesting because they are collected after each individual matcher test - if you run the same matcher over and over again you will of course hit the cache well and not have wasted data fetches.
Every benchmark case that hits the bucket has the same stats:

16206 get-range-bytes-fetched/op
0 get-range-bytes-refetched/op
61.00 get-range-bytes-requested/op
2.000 get-range-hits/op
2.000 get-range-ops/op

What we would want to is run a bunch of different matchers and then collect the caching bucket stats after they all run.
The suite of interesting test cases are seriesSelectionTestCases, for the higher level BenchmarkBucketIndexReader_ExpandedPostings.
Will need to bring these down to this area as well as fiddle with the test cache settings.

[francoposa]

Further things to test after getting the more complete benchmark setup:

• How does having a more dense sparse index-header impact object storage calls?