Skip to content

Data Retention and Privacy Policy

This document defines data retention periods and privacy/compliance alignment for CepatEdge. It is written for institutional review (e.g. university IT, data protection officers) and for operational clarity. Technical controls are described in Security Overview.


1. Scope

  • In scope: User accounts, maintenance requests, attachments, sessions, audit-related data, backups, logs.
  • Out of scope: Third-party services (e.g. email provider) — governed by their policies; we only retain what we send (e.g. delivery logs if the provider supplies them).

2. Data Retention (Default Policy)

Data categoryRetention periodRationale / notes
User account dataAccount lifetime + 30 daysAfter deletion/offboarding, 30-day grace for restore requests and audit.
Maintenance requestsIndefinite (or per institution)Core business data; can be shortened by institution (e.g. archive after N years).
Attachments (R2)Same as related requestDeleted when request is deleted or per retention override.
SessionsPer user/timeout config; max 30 daysServer-side session storage (Durable Objects); automatic cleanup.
Audit / security logs7 years (compliance-oriented)Aligns with common institutional and regulatory expectations.
Application/error logs30–90 days (configurable)Operational debugging; reduce to 30 days if no longer needed.
Database backups30 days (Neon default)Point-in-time recovery window; see Runbook for R2 backup copy.
R2 backup copiesPer backup strategy (e.g. 30–90 days)If backups are copied to R2 (auto or manual); see runbook.
Temporary/cache data24 hours or TTLDurable Object and cache TTLs; no long-term retention.

Institution override: The deploying institution may define shorter or longer retention for specific categories (e.g. maintenance requests, logs). Implementation may require configuration or one-off scripts; document any override in the institution’s own policy and in operational runbooks.


3. Privacy and Compliance Alignment

3.1 Principles

  • Data minimization: Only collect and retain what is necessary for the maintenance workflow, security, and compliance.
  • Purpose limitation: Data is used for system operation, workflow (create/approve/assign/complete), audit, and support — not for unrelated marketing or profiling.
  • Access control: Role-based access (RBAC); access to personal data limited by role and permission (see Security Overview).

3.2 Data subject rights (e.g. GDPR-style)

  • Access: Users can request a copy of their data (profile, their maintenance requests, their activity where logged). Provide via export or admin support process.
  • Rectification: Users can update profile/settings; corrections to request content depend on workflow (e.g. only before approval).
  • Erasure: Account deletion should trigger deletion or anonymization of user data and, where feasible, references in requests/logs; retention may still apply for audit/compliance (e.g. 7 years for security-relevant logs).
  • Portability: Export of user data in a machine-readable format (e.g. JSON) — implement as admin or self-service export if required by institution.

Document the exact process (who handles requests, SLA) in institution-facing docs (e.g. docs/university/for-institution/).

  • Contract/legitimate interest: For institutional deployment, processing is typically under contract (employment/university relationship) or legitimate interest (maintenance management, security).
  • Consent: Where required (e.g. non-essential communications), obtain and record consent; document in privacy notice or institution policy.

3.4 Breach notification

  • 72-hour notification: If the institution is in a jurisdiction requiring breach notification (e.g. GDPR), define internal process to assess and notify within the required window.
  • Internal process: Detection → assessment → containment → notification (see Runbook and Security Overview – Incident Response).

3.5 Vendor/subprocessor alignment

  • Cloudflare (Workers, Pages, R2, Durable Objects): See Cloudflare’s DPA and privacy documentation.
  • Neon (PostgreSQL): See Neon’s DPA and data residency options.
  • Email provider: If used for notifications, ensure DPA and retention align with institution policy.

Keep a short subprocessor list (name, purpose, location, link to DPA) for institutional audits.


4. Operational Notes

  • Backups: Backups (Neon, and any R2 copies) are part of retention; ensure backup retention and encryption match this policy (see Runbook).
  • Deletion in practice: User deletion should cascade to sessions and, where policy allows, to anonymization of user references in requests/logs; attachments and request metadata may be retained per retention table for audit.
  • Review: Review this policy periodically and when the institution’s or regulator’s requirements change.

5. References