nessar
878da155bd
- Created a pre-commit hook in .husky to run lint-staged. - Added .lintstagedrc.mjs to format frontend files with Prettier and apply Spotless to backend Java files. - Introduced .prettierignore to exclude specific directories from formatting.
3.6 KiB
3.6 KiB
GDPR Compliance and Data Archiving Strategy
Status: accepted
The application manages sensitive tenant files (identity cards, tax notices, payslips) and property condition photos. To comply with GDPR requirements and maintain a sustainable disk storage footprint over time, we have established a strict retention and image optimization policy.
Decisions
1. Automated Server-Side Image Optimization
All uploaded photos (for property condition reports) are compressed and resized on the backend before write:
- Target resolution: maximum 1920x1080 pixels.
- Format: converted to WebP.
- Compression quality factor: 75-80%. This reduces average photo size from ~6.0 MB (raw smartphone camera) to ~250 KB (95.8% reduction).
2. GDPR Retention and Purging Cycle
- Rejected Applicants: Complete deletion of all submitted documents (payslips, identity documents) exactly 30 days after rejection.
- Archived Leases:
- Upon lease termination, the lease transitions to an
Archivedstate. - Personal identity documents and income proofs of the former tenants are permanently purged 3 years after lease termination (matching the 3-year French statute of limitations for residential tenancy disputes).
- General inspection photos are purged 5 years after lease termination.
- Basic accounting records (rent ledger, invoice amounts) are retained for 10 years to comply with commercial and tax record retention laws (Article L123-22 of the French Commercial Code).
- Upon lease termination, the lease transitions to an
Storage Projections (3-Apartment Portfolio over 10 Years)
Assuming an average lease turnover rate of 2 years (i.e., 1.5 lease cycles per year across a 3-apartment portfolio).
Unit Data Footprints
| Data Type | Unoptimized Size | Optimized Size | Purge Trigger |
|---|---|---|---|
| MongoDB Text Records | ~10 KB / year | ~10 KB / year | Never (10-year archive) |
| Tenant Personal Dossier | ~5.0 MB (PDF/Images) | ~5.0 MB | Purged 3 years after lease end |
| Inspection Photos (In/Out) | ~360 MB (60 raw photos) | ~15 MB (WebP 1080p) | Purged 5 years after lease end |
| Invoices & Receipts / year | ~20 MB (Raw scans) | ~5 MB (Compressed PDFs) | Never (10-year archive) |
10-Year Cumulative Footprint Comparison
Scenario A: Unoptimized & No Purge (Default Strategy)
- Year 1: 3 * (22 MB tenant docs + 360 MB photos) + 67 MB operational files = 1,213 MB (~1.2 GB).
- Yearly Growth: 67 MB (ops) + 1.5 * (22 MB tenant docs + 720 MB in/out photos) = 1,180 MB / year.
- Year 10 Cumulative Footprint: ~11.8 GB (Continuous, compounding growth).
Scenario B: Optimized & Purged (Our Design)
- Year 1: 3 * (5 MB tenant docs + 15 MB photos) + 15 MB operational files = 75 MB.
- Yearly Growth (Before Purges): 15 MB (ops) + 1.5 * (5 MB tenant docs + 30 MB in/out photos) = 67.5 MB / year.
- Purging Offsets:
- From Year 4 onwards: Purge of tenant dossiers of expired leases saves 7.5 MB / year (1.5 leases * 5 MB).
- From Year 6 onwards: Purge of inspection photos of expired leases saves 45.0 MB / year (1.5 leases * 30 MB).
- Year 10 Cumulative Footprint: ~345 MB (Stabilizes and plates out around 350 MB due to continuous purges).
This architecture saves over 97% of server disk space over a 10-year lifecycle compared to a naive storage strategy, while strictly respecting French GDPR constraints.