# First 90 days — operations plan

A practical template for the **launch window** after first production tenant (or pilot cohort). Adjust dates to your program.

---

## Day 0–30 — Stabilization

**Staffing expectations**

- Dedicated **on-call** rotation (no “best effort”).
- CS coverage aligned with pilot time zones.
- Engineering: **≥1** person can deploy and **≥1** reviewer for production PRs.

**Monitoring intensity**

- Daily check: queue depth, webhook DLQ, export backlog, `/up`, `ops:readiness` logged.
- Dashboard review in weekly ops review minimum; increase to **daily** if any SEV-2.

**Escalation readiness**

- On-call runbook: links to [incident-escalation](./incident-escalation-framework.md), [failure matrix](../resilience/failure-scenario-matrix.md), rollback procedure.
- Pilot contact list verified (phone escalation for SEV-1).

**Stabilization checkpoints**

- End of week 2: Pilot satisfaction check (CS); telemetry vs [SLO doc](../scalability/service-level-objectives.md).
- End of week 4: Gate review for **Limited production** ([production-readiness-gates](./production-readiness-gates.md) Gate A).

**Operational freeze windows**

- **No** discretionary releases 48h before pilot’s first payroll cutover unless emergency.
- Freeze **database migrations** that touch payroll/attendance during agreed cutover unless rehearsed.

---

## Day 31–60 — Controlled expansion

**Staffing**

- Add secondary on-call if pilot count > 2 or regions expand.

**Monitoring**

- Implement first P0 metrics from [telemetry-expansion](./telemetry-expansion.md) if not already (queue age, export lag).

**Learning loops**

- Monthly resilience review; onboarding feedback themes to product ([customer-feedback-loops](./customer-feedback-loops.md)).

**Checkpoints**

- Week 8: Review tenant cap; webhook partner health; consider queue separation if backlog patterns warrant it.

---

## Day 61–90 — Confidence or constraint

**Outcome A (green):** Raise caps; document “GA-ready” risks in [final resiliency review](../resilience/final-resiliency-review.md).

**Outcome B (amber):** Maintain cap; invest in one reliability theme (webhooks, exports, or payroll observability).

**Outcome C (red):** Formal pause on new tenants; executive readout with data from monitoring + support tags.

**Quarterly drill**

- Backup restore or table-top queue outage per [operational-review-cadence](./operational-review-cadence.md).

---

## References

- `docs/rollout/production-rollout-strategy.md`
- `docs/rollout/production-monitoring.md`
- `docs/scalability/infrastructure-readiness.md`