filter: prune StoreAPI fan-out for enumerable regex __name__ matchers

[HarnageaGabriel]

• I added CHANGELOG entry for this change.

Changes

Extends the Cuckoo metric-name store filter (--enable-feature=metric-names-filter, added in #7787) so it can prune fan-out for more queries.

Before, CuckooMetricNameStoreFilter.Matches only skipped a store when the request carried an exact __name__="x" matcher. A regex matcher such as __name__=~"a|b|c" — extremely common in Grafana dashboards, recording-rule groups and or queries — was ignored, so the querier always fanned out even when none of a, b, c exist on that store.

Matches now:

• collects every __name__ matcher instead of returning on the first one;
• treats a MatchRegexp matcher whose value set is fully enumerable (labels.Matcher.SetMatches()) as a constraint, using its literal alternatives;
• keeps ignoring negative matchers and non-reducible regexes (e.g. __name__=~"a.*"), so those stay non-constraining and correctness is unchanged;
• returns false only when, for at least one constraint, none of its candidate metric names are present in the filter (AND semantics across constraints, OR within a set).

This only tightens an already-probabilistic, false-positive-only check, so it cannot cause missed data: an enumerable regex that would previously fan out unconditionally now fans out iff at least one of its names might be present.

Touched files: pkg/filter/cuckoo.go, pkg/filter/cuckoo_test.go (new), CHANGELOG.md. No proto or API changes.

Part of #1611.

Verification

• New table-driven TestCuckooMetricNameStoreFilter_Matches covering exact hit/miss, enumerable regex (all-miss vs partial-hit), non-enumerable regex, no __name__ matcher, empty matchers, multi-constraint AND, and non-__name__ regex being ignored.
• New BenchmarkCuckooMetricNameStoreFilter_Matches (equal ~56 ns/op, regexp_set ~106 ns/op on a 10k-entry filter).
• gofmt -l, go build ./pkg/filter/..., go vet ./pkg/filter/..., go test ./pkg/filter/... all pass.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 64.65%. Comparing base (8908023) to head (e2cdc22).
⚠️ Report is 33 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #9006      +/-   ##
==========================================
+ Coverage   64.51%   64.65%   +0.14%     
==========================================
  Files         289      289              
  Lines       37352    37387      +35     
==========================================
+ Hits        24097    24174      +77     
+ Misses      11158    11128      -30     
+ Partials     2097     2085      -12     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[GiedriusS]

Awesome change! Only thing I'd ask you to do is maybe we can pool the slice here? This is a very hot path