Readiness
:::info Source
Sourced from services/analytics-service/SERVICE_READINESS.md in the documentation repo.
:::
1. Level per Milestone
| M | Level | Scope |
|---|---|---|
| M0 | L1 | Event pipeline skeleton |
| M1 | L2 | Basic learner dashboards |
| M2 | L3 | Revenue dashboards |
| M3 | L3 | Compliance reports |
| M4 | L4 | Full dashboards + cohorts |
| M5 | L4 | AI insights v2, predictive |
2. Gates
G1 Domain
- Event, MetricDefinition, Dashboard, Report, ExportJob, CohortDefinition.
- 95% / 80% mutation.
G2 API
- OpenAPI + Pact + cursor + problem+json.
G3 Events
- Consumed events schemas tracked in registry.
- Published events v1.
G4 Sync
- Analytics is online-only; no sync.
G5 AI
- AIClient port for NL query, anomaly, at-risk.
- At-risk predictor: high-risk classification; HITL; bias monitoring.
G6 Observability
- Ingestion lag, query perf, export dashboards.
G7 Performance
- Ingest 100k events/sec.
- Dashboard p95 < 2s.
- Query p95 < 1s.
G8 Security
- Cross-tenant query prevention.
- PII redaction in hot tier.
- Audit-tier restricted access.
- Two-tenant iso green.
3. SLOs
Ingest lag p99 < 30s · Dashboard p95 < 2s · Export start p95 < 10s · 99.95% availability.
4. DoD
Tests + schema registry + OpenAPI + iso + mutation.
5. Release
S0 (M0)
- Event pipeline skeleton.
- NATS → Kafka → ClickHouse round-trip.
S1 (M1)
- Basic learner dashboards.
- Course completion funnel.
S4 (M3)
- Compliance reports (GDPR export, audit slice).
- Scheduled reports.
S6 (M5)
- AI insights v2 (NL query, anomaly, forecast).
- At-risk predictor with bias monitoring.
- Auto-dashboard generation.