Business Continuity / Disaster Recovery
Limited operational footprint = limited operational risk.
The architecture matters
Studio is an installable client. There is no PeerAI-side production data plane that holds customer code, prompts, or model outputs. That sharply narrows what BC/DR applies to.
The customer's BC/DR for Studio data is the customer's endpoint backup + re-install policy. PeerAI's BC/DR covers our operational services only.
PeerAI-side BC/DR scope
- Release hosting
Primary: Google Cloud Storage. Secondary: GitHub Releases. Both contain identical artefacts; either can serve customers if the other is down. RTO: 4 hours.
- Trust Center
Hosted on Vercel global edge. Static + ISR; no database. Outage scope is rendering only — content is in git and on GCS. RTO: 30 minutes.
- Marketplace / Hub
Hosted on Vercel with MongoDB Atlas. Daily backups; multi-region replica. RTO: 4 hours.
- Licensing service
Hosted on cloud infrastructure with multi-AZ redundancy. License activation can fall back to grace-period validation if the central service is briefly unavailable. RTO: 4 hours.
- Security disclosure inbox
security@peerislands.com routes to multiple operators; failover to PagerDuty for high-severity reports.
Customer-side considerations
- Studio data
Lives at ~/.peerai/ on the customer endpoint. Customer's existing endpoint backup policy applies.
- LLM provider
Customer's chosen provider's BC/DR applies. Studio supports rapid provider swap via Settings.
- Customer database
Customer's database BC/DR applies.