enrollsvc p95 4.10 s vs a normal 312 ms ± 189 ms for Mon 16:00 UTC (3 of the last 4 checks)
info
Latency anomaly
master
resolved
impact 13
Owner: Suresh Boddapati (backup: Sambasiva Bandarupalli)
opened 70d ago · last observed 2d ago · observed 22× · #1314 · b45210c3456a50730c3e
What happened template
enrollsvc's p95 is 4.1 s on master, above a normal of 312 ms ± 189 ms for this hour of the week (3 of the last 4 five-minute buckets, z = 20.0, 4 weeks of history). automated changed enrollsvc 34 minutes earlier — adjacency, not proof. Whether users noticed depends on which endpoints carried it — the service page shows the split. A p95 that stays under a second is never reported — below that nobody notices, whatever the deviation.
p95 far above its hour-of-week normal — and at least 1 s; under a second is noise.
The facts
| measure | api_p95_ms |
| unit | ms |
| value | 4,100 |
| expected_median | 312.50 |
| expected_mad | 189.03 |
| z | 20.04 |
| rel | 12.12 |
| buckets_hit | 3 |
| n | 32 |
| weeks | 4 |
| slot | Mon 16:00 UTC |
| bucket | 2026-11-16 16:00:00 UTC (2d ago) |
| floor_ms | 1,000 |
| discriminator | api_p95_ms |
| window | 2026-11-16 15:45:00 UTC → 2026-11-16 16:05:00 UTC |
Adjacent deploys — adjacency, not proof
| automated | enrollsvc · 34 min before |
History
- 2026-11-16 16:14:38 UTCreopened x-ray · severity=info was=info title=enrollsvc p95 4.10 s vs a normal 312 ms ± 189 ms for Mon 16:00 UTC (3 of the last 4 checks) level=low was_level=low finding={'id': 1314, 'fingerprint': 'b45210c3456a50730c3e', 'kind': 'anomaly.latency', 'severity': 'info', 'cluster': 'master', 'subject': 'enrollsvc', 'title': 'enrollsvc p95 4.10 s vs a normal 312 ms ± 189 ms for Mon 16:00 UTC (3 of the last 4 checks)', 'facts': {'measure': 'api_p95_ms', 'unit': 'ms', 'value': 4099.999999999998, 'expected_median': 312.5, 'expected_mad': 189.0315, 'z': 20.03634314915767, 'rel': 12.119999999999994, 'buckets_hit': 3, 'n': 32, 'weeks': 4, 'slot': 'Mon 16:00 UTC', 'bucket': 1794844800, 'window': [1794843900, 1794845100], 'floor_ms': 1000, 'recent_deploys': [{'who': 'automated', 'resource': 'enrollsvc', 'image': None, 'revision': '29620eb7090eebd0ca7163840ec0fdd716835b19', 'ts': 1794843608, 'minutes_before': 34}], 'discriminator': 'api_p95_ms'}, 'evidence': [{'label': 'Service', 'href': '/service/enrollsvc?cluster=master&from=1794822300&to=1794845100'}, {'label': 'Fleet vs normal', 'href': '/fleet?cluster=master&from=1794822300&to=1794845100'}], 'provenance': {'source': 'rollup', 'collected_at': 1794845605, 'age_seconds': 84, 'stale': False, 'shown': None, 'total': None, 'truncated': False, 'coverage_from': None, 'retain_days': None, 'note': None}, 'impact': 8.2, 'status': 'open', 'first_seen': 1788935620, 'last_seen': 1794845678, 'resolved_at': None, 'acked_by': None, 'acked_at': None, 'snoozed_until': None, 'occurrences': 22, 'narrative': "enrollsvc's p95 is 4.7 s on master, above a normal of 192 ms ± 52 ms for this hour of the week (3 of the last 4 five-minute buckets, z = 86.9, 4 weeks of history). Whether users noticed depends on which endpoints carried it — the service page shows the split. A p95 that stays under a second is never reported — below that nobody notices, whatever the deviation.", 'narrative_kind': 'template', 'discriminator': 'api_p95_ms', 'level': 'low', 'told_level': None, 'episode_at': 1794845678, 'chat_thread': None, 'kind_label': 'Latency anomaly', 'group': 'Anomalies', 'age_s': 5910069, 'stale_s': 11}
- 2026-11-16 16:32:01 UTCresolved x-ray · reason=not observed grace_s=900 finding={'id': 1314, 'fingerprint': 'b45210c3456a50730c3e', 'kind': 'anomaly.latency', 'severity': 'info', 'cluster': 'master', 'subject': 'enrollsvc', 'title': 'enrollsvc p95 4.10 s vs a normal 312 ms ± 189 ms for Mon 16:00 UTC (3 of the last 4 checks)', 'facts': {'measure': 'api_p95_ms', 'unit': 'ms', 'value': 4099.999999999998, 'expected_median': 312.5, 'expected_mad': 189.0315, 'z': 20.03634314915767, 'rel': 12.119999999999994, 'buckets_hit': 3, 'n': 32, 'weeks': 4, 'slot': 'Mon 16:00 UTC', 'bucket': 1794844800, 'window': [1794843900, 1794845100], 'floor_ms': 1000, 'recent_deploys': [{'who': 'automated', 'resource': 'enrollsvc', 'image': None, 'revision': '29620eb7090eebd0ca7163840ec0fdd716835b19', 'ts': 1794843608, 'minutes_before': 34}], 'discriminator': 'api_p95_ms'}, 'evidence': [{'label': 'Service', 'href': '/service/enrollsvc?cluster=master&from=1794822300&to=1794845100'}, {'label': 'Fleet vs normal', 'href': '/fleet?cluster=master&from=1794822300&to=1794845100'}], 'provenance': {'source': 'rollup', 'collected_at': 1794845605, 'age_seconds': 84, 'stale': False, 'shown': None, 'total': None, 'truncated': False, 'coverage_from': None, 'retain_days': None, 'note': None}, 'impact': 13.2, 'status': 'resolved', 'first_seen': 1788935620, 'last_seen': 1794845678, 'resolved_at': 1794846721, 'acked_by': None, 'acked_at': None, 'snoozed_until': None, 'occurrences': 22, 'narrative': "enrollsvc's p95 is 4.1 s on master, above a normal of 312 ms ± 189 ms for this hour of the week (3 of the last 4 five-minute buckets, z = 20.0, 4 weeks of history). automated changed enrollsvc 34 minutes earlier — adjacency, not proof. Whether users noticed depends on which endpoints carried it — the service page shows the split. A p95 that stays under a second is never reported — below that nobody notices, whatever the deviation.", 'narrative_kind': 'template', 'discriminator': 'api_p95_ms', 'level': 'low', 'told_level': None, 'episode_at': 1794845678, 'chat_thread': None, 'kind_label': 'Latency anomaly', 'group': 'Anomalies', 'age_s': 5911112, 'stale_s': 1054}
Check it yourself
Every link carries the finding's own time window, so it shows tomorrow what it shows now.
Actions
Sign in (IAP) to acknowledge, snooze or resolve.
Where the numbers come from
| source | rollup |
| collected_at | 1794845605 |
| age_seconds | 84 |
| stale | False |
| shown | None |
| total | None |
| truncated | False |
| coverage_from | None |
| retain_days | None |
| note | None |
| detector | seasonal |