What this PR does:

TestCompactor_ShouldCompactAllUsersOnShardingEnabledButOnlyOneInstanceRunning fails intermittently on the arm64 -race CI job: the background userIndexUpdateLoop ticks every ~100ms in tests (via prepare()'s user-index strategy) and logs a ring error whenever no healthy instance exists — before the lifecycler is ACTIVE or during LEAVING at shutdown. Any tick landing in the test window leaks an extra line into the exact log-set assertion, failing with elements differ.

removeIgnoredLogs already suppresses two of the loop's log lines via map entries; the two error lines were missed because their variable err= payloads require regexes. This adds them:

ignoredLogStringsRegexList := []*regexp.Regexp{
	regexp.MustCompile(`^level=(info|debug|warn) component=cleaner .+$`),
	regexp.MustCompile(`^level=info component=compactor msg="set state" .+$`),
	regexp.MustCompile(`^level=error component=compactor msg="failed to check if compactor owns updating user index" err=.+$`),
	regexp.MustCompile(`^level=error component=compactor msg="failed to update user index" err=.+$`),
}

Since all 14 exact log-set assertions in compactor_test.go and compactor_paritioning_test.go funnel through removeIgnoredLogs, this single change covers every sharding-enabled test exposed to the same race. Regexes were verified against real go-kit logfmt output, including the exact err="at least 1 live replicas required, could only find 0" line from the failed CI jobs. Test-only change; no production behavior modified.

Checklist

• Tests updated
• Documentation added
• CHANGELOG.md updated - the order of entries should be [CHANGE], [FEATURE], [ENHANCEMENT], [BUGFIX]
• docs/configuration/v1-guarantees.md updated if this PR introduces experimental flags

• Fixes Flaky test: TestCompactor_ShouldCompactAllUsersOnShardingEnabledButOnlyOneInstanceRunning (arm64) — userIndexUpdateLoop ring errors leak into the exact log-set assertion #7826