[tempodb]: remove dead ring-based compaction code

[zalegrala]

What this PR does:

Removes the dead code left behind after the ring-based compaction path was superseded by the backend scheduler/worker architecture. Since compactors no longer run as separate ring members, the following were unreachable from any live code path and are deleted:

• EnableCompaction (the entry point for the ring-based path)
• compactionLoop / compactOneTenant / compactWhileOwns / compactOneJob / doForAtLeast
• retentionLoop / doRetention
• CompactorConfig fields: MaxTimePerTenant, CompactionCycle (including their defaults/flag registration)
• readerWriter struct fields: compactorCfg, compactorSharder, compactorOverrides, compactorTenantOffset

CompactWithConfig and RetainWithConfig (called directly by the backend worker) are the only remaining compaction/retention paths and are kept.

Retention tests are refactored to call RetainWithConfig directly with explicit config parameters rather than relying on the now-deleted EnableCompaction setup + doRetention pattern.

Which issue(s) this PR fixes:
Fixes #

Checklist

• Tests updated
• Documentation added
• CHANGELOG.md updated - the order of entries should be [CHANGE], [FEATURE], [ENHANCEMENT], [BUGFIX]

[Copilot]

Pull request overview

Removes the legacy ring-based compaction/retention entrypoints and loops from tempodb now that compaction/retention is driven by the backend scheduler/worker path (CompactWithConfig / RetainWithConfig). Updates tests to stop relying on the deleted setup path.

Changes:

• Deletes EnableCompaction and the ring-loop based compaction/retention codepaths.
• Removes CompactorConfig fields that only applied to the deleted loops (max_time_per_tenant, compaction_cycle) and refactors tests to call CompactWithConfig / RetainWithConfig directly.
• Simplifies/updates several tests that previously enabled compaction just to exercise polling/search behavior.

Reviewed changes

Copilot reviewed 10 out of 10 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
tempodb/tempodb.go	Removes EnableCompaction from the Compactor interface and drops now-unused compactor fields from readerWriter.
tempodb/compactor.go	Deletes the ring-loop compaction orchestration helpers; leaves CompactWithConfig as the main entrypoint.
tempodb/retention.go	Removes the old retention loop wrapper, leaving RetainWithConfig / RetainTenantWithConfig.
tempodb/config.go	Removes defaults/fields related to the deleted loop path (max_time_per_tenant, compaction_cycle) and drops CompactorConfig.validate().
tempodb/config_test.go	Removes the unit test that only validated the deleted CompactorConfig.validate() method.
tempodb/tempodb_test.go	Removes EnableCompaction usage in tests; updates one test to compact via CompactWithConfig.
tempodb/tempodb_search_test.go	Removes EnableCompaction setup from search test harnesses.
tempodb/tempodb_metrics_test.go	Removes EnableCompaction setup from metrics query test harness.
tempodb/retention_test.go	Refactors retention tests to call RetainWithConfig directly with explicit config.
tempodb/compactor_test.go	Removes tests/benchmarks that depended on the deleted ring-loop compaction code; keeps direct compaction tests.

Comments suppressed due to low confidence (1)

tempodb/retention.go:18

• CRITICAL: RetainWithConfig will panic if compactorCfg is nil or if RetentionConcurrency is 0 (because boundedwaitgroup.New(0) panics). Previously EnableCompaction defaulted RetentionConcurrency for tests, but now callers can more easily hit this with misconfiguration (e.g. retention_concurrency: 0). Can we defensively default to DefaultRetentionConcurrency (or return early with a logged error) before constructing the bounded waitgroup?

func (rw *readerWriter) RetainWithConfig(ctx context.Context, compactorCfg *CompactorConfig, compactorSharder CompactorSharder, compactorOverrides CompactorOverrides) {
	tenants := rw.blocklist.Tenants()

	bg := boundedwaitgroup.New(compactorCfg.RetentionConcurrency)

[Copilot]

Pull request overview

Copilot reviewed 13 out of 13 changed files in this pull request and generated 4 comments.