Run twenty client websites under one Agency account. Each as separate company.website — your team holds the keys, your clients see only their site.
- 20 websites, 20 custom domains, 20 res.company records
- Dedicated Odoo DB + full white-label admin
- 120 GB storage · ~1.5M page-views / month
- Sub-account roles per website (designer, editor, billing)
- Centralised billing — invoice your clients direct, we invoice the Agency once
- Nightly database dump at 02:30 and a separate off-site copy at 03:30 held with a different provider; write-ahead-log archiving on the primary gives local point-in-time recovery covering the period since 23 July 2026, 15:03. Until 24 July 2026 the dump covered the database only — attachments and uploaded documents were not included; the fix is deployed and the first backup containing documents is the run of the night of 24–25 July 2026
- Retention target 30 daily / 4 weekly / 6 monthly / 2 yearly locally, 14 daily / 8 weekly / 12 monthly / 2 yearly off-site — with the copies for 21, 22 and 23 July 2026 missing after a disk-full incident, so the daily target has a three-day gap. Backups travel over TLS and are not encrypted at rest
- Fleet restore drill on a weekly timer, Sundays 03:30 — client databases restored into throwaway copies, database records only, not one drill per customer; every result published pass or fail, with the dates and figures on the proof page. The first drill was started by hand on 24 July 2026 and passed. A separate full-stack drill, Sundays 05:00, restores the whole cluster and boots a tenant; its first scheduled run failed and its attempts are published too
- No uptime percentage published — we publish the restore drill instead
- E-mail support, one working day, from the engineer who built it — 6 engineering hours a month