Data & Sensing
Collect less, secure the rest. Treat telemetry as signed evidence with provenance; process locally when it matters.
Healthy uses
- Local buffering and signed telemetry at the edge.
- Strict retention policies and data minimization.
- Reproducible pipelines with lineage and quality gates.
Risks to avoid
- PII in logs; uncontrolled data sprawl.
- Vendor‑locked schemas without contracts.
- No offline path when networks drop.
Controls & guardrails
- Data classification; access tiers; quarterly reviews.
- Data contracts + schema registry; breaking‑change policy.
- SLOs: ingest latency, missing‑data rate, recovery time.
- Provenance/signing (Sigstore); SBOM for data tools.
- DR playbooks: replays, backfills, offline sync windows.
Signals & metrics
- Schema‑compatibility score
- Data‑quality test pass rate
- Backfill debt
Tech hooks
Edge/IoT, Integration, Data Platform, Security (Zero Trust), Observability.
Related playbooks
Point A → Point B, Fast
Architecture Sprint (1–2 weeks): Current → Target → 1–90-day plan. D1 maps, D2 target, D3 plan, D4 exec brief.