Staying up: DR and multi-region

Running it yourself means owning availability, and the platform is built to make that tractable rather than promising a number it has not earned. Key points:

• The platform is designed for high availability with disaster-recovery and multi-region Helm profiles. These are deployment profiles you can run, not a measured service-level guarantee.
• Be careful with claims here. Do not state an achieved uptime figure. The honest framing is "designed for high availability," because the profiles give you the shape, and your own operation produces the actual numbers.
• Tenant isolation still holds in your install: tenants are kept apart at the database with row-level security, so the separation does not depend on application code being bug-free, even on infrastructure you run.