Mark 03809_time64_not_monotonic as long test

[groeneai]

Adds the long tag to tests/queries/0_stateless/03809_time64_not_monotonic.sql so the test opts into the long-test budget and stops being clipped by the 600s timeout on the slowest CI lane (amd_msan, WasmEdge, ~8x slower than amd_tsan).

Follow-up to @alexey-milovidov directive on #102498 (comment) ("do a stabilization PR, option 2 with tagging the test as long").

Triage data (90 days): 195,745 runs, 10 timeouts (99.995% pass rate), p99 = 404s, max passing = 563s — pure transient timeout, no test bug. Precedents for naming/tagging slow regression tests as long: 01049_join_low_card_bug_long, 01516_drop_table_stress_long, 01443_merge_truncate_long.

One-line change to the .sql file. No test logic touched, no .reference change.

Changelog category (leave one):

• CI Fix or Improvement (changelog entry is not required)

Changelog entry (a user-readable short description of the changes that goes into CHANGELOG.md):

...

Documentation entry for user-facing changes

• Documentation is written (mandatory for new features)

[groeneai]

Pre-PR validation gate (per TASK.md Phase 4 step 9)

#	Question	Answer
a	Deterministic repro?	N/A — transient timeout, not a deterministic bug. The fix is a CI metadata change (opt the test into the long-test budget), not a code fix.
b	Root cause explained?	Yes. The test runs ~p99=404s, max passing=563s on the slowest lane (amd_msan, WasmEdge is ~8x slower than amd_tsan). With a 600s timeout there is too little headroom, so a small fraction of runs (~0.005%) hit the timeout. The long tag re-buckets it into the 1200–1800s budget.
c	Fix matches root cause?	Yes. The long tag is the established mechanism for tests whose runtime is genuinely long under the slowest sanitizer lanes; precedents include 01049_join_low_card_bug_long, 01516_drop_table_stress_long, 01443_merge_truncate_long. No assertion bounds widened, no setting tags added, no data sizes reduced.
d	Test intent preserved?	Yes. Zero changes to test logic, queries, or expected output. Only metadata (the Tags: line). The regression coverage of issue #95667 (non-monotonic Time64 conversion in MergeTreeSetIndex) is fully retained.
e	Demonstrated in both directions?	N/A for the transient-timeout case. Equivalent precedent: tests that previously timed out on the slow lanes have stopped timing out after gaining the long tag (e.g. 01443_merge_truncate_long).
f	Fix is general, not a narrow patch?	N/A — not a code fix; it's a CI tag. Scope is intentionally narrow: one test file, one line.

Session: cron:clickhouse-ci-task-worker:20260429-134500

[groeneai]

cc @alexey-milovidov — could you review this? It's the stabilization PR you asked for on #102498: tags 03809_time64_not_monotonic as long so the test gets the long-test budget on the slowest CI lane (amd_msan, WasmEdge, ~8x slower than amd_tsan). One-line metadata change, no test logic touched.

[clickhouse-gh]

Workflow [PR], commit [815fff6]

Summary: ❌

job_name	test_name	status	info	comment
Stateless tests (arm_asan_ubsan, azure, parallel)		FAIL
	03918_json_lazy_type_hints_merge	FAIL	cidb
	01006_simpod_empty_part_single_column_write	FAIL	cidb
	02971_analyzer_remote_id	FAIL	cidb
	02706_kolmogorov_smirnov_test_scipy	FAIL	cidb
	01765_move_to_table_overlapping_block_number	FAIL	cidb
	04071_global_in_dia_no_explicit_set_elements	FAIL	cidb
	03754_point_in_polygon_primary_key_index	FAIL	cidb
	01889_check_row_policy_defined_using_user_function	FAIL	cidb
	03367_l2_distance_transposed_2	FAIL	cidb
	03781_json_max_dynamic_subcolumns_control_on_parsing	FAIL	cidb
	31 more test cases not shown

---

AI Review

Summary

This PR adds -- Tags: long to tests/queries/0_stateless/03809_time64_not_monotonic.sql so the test runs with the long-test budget. The change is minimal, scoped to test scheduling metadata, and does not alter test logic or product behavior; no blockers or majors were found.

ClickHouse Rules

Item	Status	Notes
Deletion logging	➖
Serialization versioning	➖
Core-area scrutiny	➖
No test removal	✅
Experimental gate	➖
No magic constants	✅
Backward compatibility	✅
SettingsChangesHistory.cpp	➖
PR metadata quality	✅
Safe rollout	✅
Compilation time	➖
No large/binary files	✅

Final Verdict

• Status: ✅ Approve