docsvc p95 4.47 s vs a normal 91 ms ± 63 ms for Sat 18:00 UTC (3 of the last 4 checks)
info
Latency anomaly
test27
resolved
impact 10
Owner: Prasad Limaye (backup: Harish Kumar)
opened 68d ago · last observed 4d ago · observed 69× · #1499 · 5da3bb23aeb66acd3080
What happened template
docsvc's p95 is 4.5 s on test27, above a normal of 91 ms ± 63 ms for this hour of the week (3 of the last 4 five-minute buckets, z = 69.2, 2 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 | 4,475 |
| expected_median | 91.25 |
| expected_mad | 63.38 |
| z | 69.16 |
| rel | 48.04 |
| buckets_hit | 3 |
| n | 24 |
| weeks | 2 |
| slot | Sat 18:00 UTC |
| bucket | 2026-11-14 18:15:00 UTC (4d ago) |
| floor_ms | 1,000 |
| discriminator | api_p95_ms |
| window | 2026-11-14 18:00:00 UTC → 2026-11-14 18:20:00 UTC |
What is failing in docsvc — the bad events, decomposed
“Bad events” are HTTP 5xx answers from docsvc. In this finding's window (2026-11-14 18:00:00 UTC → 2026-11-14 18:20:00 UTC) there were 0; every line links to the grouped errors behind it.
The error shapes behind them
| 9× | Request completed — 500 · could not update document metadata for doc id 7889000a-05cd-461c-b2ba-6754ef50a40d: could not … since 2026-10-01 01:01:09 UTC |
| 4× | Request completed — 500 · document is already in Staging collection, no promotion needed since 2026-11-03 16:40:25 UTC |
| 3× | deliverable '01.01.01 Trial Master File Plan': Mismatch: locationId '01.01.01 Trial Master File Plan' does not match res… since 2026-11-16 18:59:33 UTC |
| 2× | failed to process db result — Neo4jError: Neo.TransientError.Transaction.DeadlockDetected (ForsetiClient[transactionId=1… since 2026-11-03 12:36:16 UTC |
| 2× | Request completed — 500 · signer role \ since 2026-11-16 11:32:22 UTC |
History
- 2026-11-14 18:29:46 UTCreopened x-ray · severity=info was=info title=docsvc p95 4.47 s vs a normal 91 ms ± 63 ms for Sat 18:00 UTC (3 of the last 4 checks) level=low was_level=low finding={'id': 1499, 'fingerprint': '5da3bb23aeb66acd3080', 'kind': 'anomaly.latency', 'severity': 'info', 'cluster': 'test27', 'subject': 'docsvc', 'title': 'docsvc p95 4.47 s vs a normal 91 ms ± 63 ms for Sat 18:00 UTC (3 of the last 4 checks)', 'facts': {'measure': 'api_p95_ms', 'unit': 'ms', 'value': 4474.999999999999, 'expected_median': 91.25, 'expected_mad': 63.38115, 'z': 69.16488577439821, 'rel': 48.04109589041095, 'buckets_hit': 3, 'n': 24, 'weeks': 2, 'slot': 'Sat 18:00 UTC', 'bucket': 1794680100, 'window': [1794679200, 1794680400], 'floor_ms': 1000, 'recent_deploys': [], 'discriminator': 'api_p95_ms'}, 'evidence': [{'label': 'Service', 'href': '/service/docsvc?cluster=test27&from=1794657600&to=1794680400'}, {'label': 'Fleet vs normal', 'href': '/fleet?cluster=test27&from=1794657600&to=1794680400'}], 'provenance': {'source': 'rollup', 'collected_at': 1794680985, 'age_seconds': 11, 'stale': False, 'shown': None, 'total': None, 'truncated': False, 'coverage_from': None, 'retain_days': None, 'note': None}, 'impact': 5.1, 'status': 'open', 'first_seen': 1789144123, 'last_seen': 1794680986, 'resolved_at': None, 'acked_by': None, 'acked_at': None, 'snoozed_until': None, 'occurrences': 69, 'narrative': "docsvc's p95 is 2.4 s on test27, above a normal of 98 ms ± 27 ms for this hour of the week (3 of the last 4 five-minute buckets, z = 88.6, 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': 1794680986, 'chat_thread': None, 'kind_label': 'Latency anomaly', 'group': 'Anomalies', 'age_s': 5536874, 'stale_s': 11}
- 2026-11-16 06:49:27 UTCresolved x-ray · reason=not observed grace_s=900 finding={'id': 1499, 'fingerprint': '5da3bb23aeb66acd3080', 'kind': 'anomaly.latency', 'severity': 'info', 'cluster': 'test27', 'subject': 'docsvc', 'title': 'docsvc p95 4.47 s vs a normal 91 ms ± 63 ms for Sat 18:00 UTC (3 of the last 4 checks)', 'facts': {'measure': 'api_p95_ms', 'unit': 'ms', 'value': 4474.999999999999, 'expected_median': 91.25, 'expected_mad': 63.38115, 'z': 69.16488577439821, 'rel': 48.04109589041095, 'buckets_hit': 3, 'n': 24, 'weeks': 2, 'slot': 'Sat 18:00 UTC', 'bucket': 1794680100, 'window': [1794679200, 1794680400], 'floor_ms': 1000, 'recent_deploys': [], 'discriminator': 'api_p95_ms'}, 'evidence': [{'label': 'Service', 'href': '/service/docsvc?cluster=test27&from=1794657600&to=1794680400'}, {'label': 'Fleet vs normal', 'href': '/fleet?cluster=test27&from=1794657600&to=1794680400'}], 'provenance': {'source': 'rollup', 'collected_at': 1794680985, 'age_seconds': 11, 'stale': False, 'shown': None, 'total': None, 'truncated': False, 'coverage_from': None, 'retain_days': None, 'note': None}, 'impact': 10.1, 'status': 'resolved', 'first_seen': 1789144123, 'last_seen': 1794680986, 'resolved_at': 1794811767, 'acked_by': None, 'acked_at': None, 'snoozed_until': None, 'occurrences': 69, 'narrative': "docsvc's p95 is 4.5 s on test27, above a normal of 91 ms ± 63 ms for this hour of the week (3 of the last 4 five-minute buckets, z = 69.2, 2 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': 1794680986, 'chat_thread': None, 'kind_label': 'Latency anomaly', 'group': 'Anomalies', 'age_s': 5667659, 'stale_s': 130796}
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 | 1794680985 |
| age_seconds | 11 |
| stale | False |
| shown | None |
| total | None |
| truncated | False |
| coverage_from | None |
| retain_days | None |
| note | None |
| detector | seasonal |