From 7c95d5ca57d1f3c38f8145218abbdf0f662812d1 Mon Sep 17 00:00:00 2001 From: jenkins Date: Fri, 7 Aug 2026 20:01:40 -0300 Subject: [PATCH] monitoring(testing): keep quality rollups current --- scripts/tests/test_monitoring_query_capacity.py | 5 +++-- services/monitoring/vmalert-atlas-availability.yaml | 5 ++++- 2 files changed, 7 insertions(+), 3 deletions(-) diff --git a/scripts/tests/test_monitoring_query_capacity.py b/scripts/tests/test_monitoring_query_capacity.py index 3f8126847..05b366f49 100644 --- a/scripts/tests/test_monitoring_query_capacity.py +++ b/scripts/tests/test_monitoring_query_capacity.py @@ -154,13 +154,14 @@ def test_daily_availability_backfill_stays_below_query_timeout() -> None: def test_quality_rollups_do_not_run_every_minute() -> None: - """Keep high-cardinality quality rollups below the backend saturation cadence.""" + """Keep quality rollups cheap while avoiding instant-query staleness gaps.""" manifest = _documents( REPO_ROOT / "services/monitoring/vmalert-atlas-availability.yaml" )[0] quality = yaml.safe_load(manifest["data"]["platform-quality.yaml"])["groups"][0] - assert quality["interval"] == "5m" + assert quality["interval"] == "4m" + assert quality["eval_offset"] == "1m" def test_vmalert_reloads_updated_rule_files() -> None: diff --git a/services/monitoring/vmalert-atlas-availability.yaml b/services/monitoring/vmalert-atlas-availability.yaml index 8d8a3c3b5..0bdabcea9 100644 --- a/services/monitoring/vmalert-atlas-availability.yaml +++ b/services/monitoring/vmalert-atlas-availability.yaml @@ -99,7 +99,10 @@ data: platform-quality.yaml: | groups: - name: platform.quality - interval: 5m + # Stay inside the five-minute instant-query lookback, including the + # group's evaluation and remote-write time. + interval: 4m + eval_offset: 1m rules: - record: platform_quality:test_case_status:count_1h expr: |