Phase 72.5.2.13 adds @v2/telemetry-alerting-rules, the operational alerting
surface for crash spikes, player count drops, and matchmaking wait regressions.
The service composes @iris/analytics-realtime alert rule types through
buildV2TelemetryAlertingRulesSurface. It evaluates aggregate telemetry samples
and emits deterministic alert rows for:
crash-rate-spike-2x: crash rate multiplier greater than or equal to2player-count-drop-20pct: player count drop greater than or equal to0.2matchmaking-wait-time-3min: p95 matchmaking wait greater than or equal to180seconds
Raw account IDs are not accepted or emitted. Inputs are operational aggregates such as crash rate, baseline crash rate, current player count, baseline player count, and p95 matchmaking wait time.
Policy#
- Policy ID:
v2-telemetry-alerting-rules-v1 - Source library:
@iris/analytics-realtime - Rule families: crash rate spike, player count drop, matchmaking wait time
- Thresholds: crash rate spike
2x, player count drop20%, matchmaking wait3 minutes - Privacy: aggregate samples only; raw account IDs are never exposed
Output#
The surface emits:
- the three
AlertRuledefinitions used by operators and gates - latest-sample evaluations per rule family
- deterministic active alert rows with rule ID, severity, observed value, and threshold
- summary counts for active, critical, and warning alerts
- chart-ready copies of evaluations and alerts
Gates#
check-v2-telemetry-alerting-rules.py validates the contract, service package,
source tokens, tests, docs, workflow wiring, Horde gates, and the Phase 72
checklist item. Required gates are telemetry-alerting-rules-service-present,
telemetry-alerting-rules-crash-rate-spike,
telemetry-alerting-rules-player-count-drop,
telemetry-alerting-rules-matchmaking-wait, and
telemetry-alerting-rules-ci-wired.
Targeted verification:
pnpm --filter @v2/telemetry-alerting-rules run typecheck
pnpm --filter @v2/telemetry-alerting-rules run test
python V2/ue/Tools/check-v2-telemetry-alerting-rules.py