# Post-market Monitoring Plan Template (with sampling policy)

Not legal advice. This is a monitoring plan template designed to be auditable: it should define
signals, thresholds, owners, and an incident workflow (not just “we will monitor”).

Version: v1.0  
Last updated: 2025-12-16  
Changelog:
- 2025-12-16 v1.0 Initial release

Report an issue: https://kla.digital/contact?subject=Template%20issue%20-%20Post-market%20Monitoring%20Plan

## 0) Plan metadata

- Plan owner:
- Approvers:
- Effective date:
- Systems in scope:
- Review cadence (e.g., monthly):

## 1) Scope & boundaries

- [ ] Intended purpose (and “do not use for”)
- [ ] Environments in scope (prod, staging)
- [ ] User groups and impacted persons
- [ ] Dependencies (models, tools, data sources)

## 2) Monitoring signals (what you measure)

Quality signals:

- [ ] Grounding/accuracy sampling (define method)
- [ ] Output completeness and format validity

Safety & policy signals:

- [ ] Policy violations (blocked) and near-misses (nearly blocked)
- [ ] Tool/action correctness (wrong system, wrong action, wrong record)

Operational health:

- [ ] Latency, throughput, cost
- [ ] Error rates and retries

## 3) Thresholds & triggers (what happens when breached)

For each signal, define:

- [ ] Threshold(s) and severity level mapping
- [ ] Owner(s) and SLA to respond
- [ ] Automated actions (block, require-review, degrade mode)
- [ ] Human actions (escalation, approval required, rollback)

## 4) Sampling policy (risk-tiered, auditable)

### 4.1 What gets sampled

- [ ] Risk tiers (low/medium/high) and how they are assigned
- [ ] Always-sample conditions (high-risk actions, new versions, incidents)

### 4.2 Sampling rates

- [ ] Baseline sampling rates per tier (e.g., high 25%, medium 5%, low 1%)
- [ ] Burst sampling rules (during incidents, after major changes)

### 4.3 Reviewer guidance

- [ ] What reviewers mark (labels, rubrics)
- [ ] How disagreements are handled (second review, adjudication)
- [ ] Training requirements and periodic calibration

## 5) Incident management

- [ ] Severity levels (P0/P1/P2) and examples
- [ ] Response SLAs
- [ ] Rollback / kill-switch procedure
- [ ] Reporting responsibilities (internal/external)

## 6) Change management (material changes)

- [ ] Define “material change” for this system
- [ ] Re-validation requirements
- [ ] Versioning and approval process
- [ ] Ramp and rollback strategy

## 7) Evidence you keep (minimum)

- [ ] Monitoring dashboards + snapshots (per cadence)
- [ ] Sampling outcomes (queue records + reviewer labels)
- [ ] Policy pack versions + enforcement logs
- [ ] Incident reports + corrective actions + verification evidence
- [ ] Change approvals and release notes