enrollsvc p95 1.45 s vs a normal 112 ms ± 44 ms for Mon 08:00 UTC (3 of the last 4 checks)
info
Latency anomaly
automation
resolved
impact 12
Owner: Suresh Boddapati (backup: Sambasiva Bandarupalli)
opened 72d ago · last observed 3d ago · observed 27× · #1186 · eef3d4ac5d77b0fd1707
What happened template
enrollsvc's p95 is 1.4 s on automation, above a normal of 112 ms ± 44 ms for this hour of the week (3 of the last 4 five-minute buckets, z = 30.3, 3 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.
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 | 1,450 |
| expected_median | 112.00 |
| expected_mad | 44.11 |
| z | 30.34 |
| rel | 11.95 |
| buckets_hit | 3 |
| n | 21 |
| weeks | 3 |
| slot | Mon 08:00 UTC |
| bucket | 2026-11-16 08:20:00 UTC (3d ago) |
| floor_ms | 1,000 |
| discriminator | api_p95_ms |
| window | 2026-11-16 08:05:00 UTC → 2026-11-16 08:25:00 UTC |
What is failing in enrollsvc — the bad events, decomposed
“Bad events” are HTTP 5xx answers from enrollsvc. In this finding's window (2026-11-16 08:05:00 UTC → 2026-11-16 08:25:00 UTC) there were 0; every line links to the grouped errors behind it.
The error shapes behind them
| 12× | Request completed — 500 · study has sites bound to it. Please delete sites before deletion since 2026-09-28 07:28:01 UTC |
| 2× | write failed — write tcp 10.129.3.90:7174->127.0.0.6:39037: write: broken pipe since 2026-10-26 03:05:06 UTC |
| 1× | Request completed — 500 · Countries "United States" have study site(s), cannot be removed since 2026-09-25 08:56:36 UTC |
History
- 2026-11-16 08:39:06 UTCreopened x-ray · severity=info was=info title=enrollsvc p95 1.45 s vs a normal 112 ms ± 44 ms for Mon 08:00 UTC (3 of the last 4 checks) level=low was_level=low finding={'id': 1186, 'fingerprint': 'eef3d4ac5d77b0fd1707', 'kind': 'anomaly.latency', 'severity': 'info', 'cluster': 'automation', 'subject': 'enrollsvc', 'title': 'enrollsvc p95 1.45 s vs a normal 112 ms ± 44 ms for Mon 08:00 UTC (3 of the last 4 checks)', 'facts': {'measure': 'api_p95_ms', 'unit': 'ms', 'value': 1449.9999999999993, 'expected_median': 112.0, 'expected_mad': 44.10735, 'z': 30.33508020772047, 'rel': 11.946428571428566, 'buckets_hit': 3, 'n': 21, 'weeks': 3, 'slot': 'Mon 08:00 UTC', 'bucket': 1794817200, 'window': [1794816300, 1794817500], 'floor_ms': 1000, 'recent_deploys': [], 'discriminator': 'api_p95_ms'}, 'evidence': [{'label': 'Service', 'href': '/service/enrollsvc?cluster=automation&from=1794794700&to=1794817500'}, {'label': 'Fleet vs normal', 'href': '/fleet?cluster=automation&from=1794794700&to=1794817500'}], 'provenance': {'source': 'rollup', 'collected_at': 1794818005, 'age_seconds': 352, 'stale': False, 'shown': None, 'total': None, 'truncated': False, 'coverage_from': None, 'retain_days': None, 'note': None}, 'impact': 7.1, 'status': 'open', 'first_seen': 1788917489, 'last_seen': 1794818346, 'resolved_at': None, 'acked_by': None, 'acked_at': None, 'snoozed_until': None, 'occurrences': 27, 'narrative': "enrollsvc's p95 is 4.7 s on automation, above a normal of 142 ms ± 64 ms for this hour of the week (3 of the last 4 five-minute buckets, z = 71.0, 3 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': 1794818346, 'chat_thread': None, 'kind_label': 'Latency anomaly', 'group': 'Anomalies', 'age_s': 5900869, 'stale_s': 12}
- 2026-11-16 09:06:49 UTCresolved x-ray · reason=not observed grace_s=900 finding={'id': 1186, 'fingerprint': 'eef3d4ac5d77b0fd1707', 'kind': 'anomaly.latency', 'severity': 'info', 'cluster': 'automation', 'subject': 'enrollsvc', 'title': 'enrollsvc p95 1.45 s vs a normal 112 ms ± 44 ms for Mon 08:00 UTC (3 of the last 4 checks)', 'facts': {'measure': 'api_p95_ms', 'unit': 'ms', 'value': 1449.9999999999993, 'expected_median': 112.0, 'expected_mad': 44.10735, 'z': 30.33508020772047, 'rel': 11.946428571428566, 'buckets_hit': 3, 'n': 21, 'weeks': 3, 'slot': 'Mon 08:00 UTC', 'bucket': 1794817200, 'window': [1794816300, 1794817500], 'floor_ms': 1000, 'recent_deploys': [], 'discriminator': 'api_p95_ms'}, 'evidence': [{'label': 'Service', 'href': '/service/enrollsvc?cluster=automation&from=1794794700&to=1794817500'}, {'label': 'Fleet vs normal', 'href': '/fleet?cluster=automation&from=1794794700&to=1794817500'}], 'provenance': {'source': 'rollup', 'collected_at': 1794818005, 'age_seconds': 352, 'stale': False, 'shown': None, 'total': None, 'truncated': False, 'coverage_from': None, 'retain_days': None, 'note': None}, 'impact': 12.1, 'status': 'resolved', 'first_seen': 1788917489, 'last_seen': 1794818346, 'resolved_at': 1794820009, 'acked_by': None, 'acked_at': None, 'snoozed_until': None, 'occurrences': 27, 'narrative': "enrollsvc's p95 is 1.4 s on automation, above a normal of 112 ms ± 44 ms for this hour of the week (3 of the last 4 five-minute buckets, z = 30.3, 3 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': 1794818346, 'chat_thread': None, 'kind_label': 'Latency anomaly', 'group': 'Anomalies', 'age_s': 5902534, 'stale_s': 1677}
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 | 1794818005 |
| age_seconds | 352 |
| stale | False |
| shown | None |
| total | None |
| truncated | False |
| coverage_from | None |
| retain_days | None |
| note | None |
| detector | seasonal |