FAIRVISORBuilt for paranoid engineers
Fairvisor is designed with a minimal trust surface. Your traffic data never touches our servers. Your edge runs on your infrastructure. Here's exactly how security works.
Data Protection
| Aspect | Detail |
|---|---|
| Data in transit | TLS 1.2+ required, mTLS optional |
| Data at rest | AES-256 encryption (SaaS database) |
| Policy bundles | HMAC-SHA256 signed, tamper-detected at edge |
| EDGE_TOKEN | Rotatable, scoped to single edge instance |
| JWT validation | Configurable algorithms (RS256, ES256), JWKS auto-refresh |
| Secrets in policies | Never stored in policy bundles — referenced, not embedded |
Access Control
| Role | Permissions |
|---|---|
| Viewer | Read policies, view analytics, view audit log |
| Editor | Create/edit policies (draft only) |
| Operator | Deploy policies, activate kill-switch, manage edges |
| Admin | Manage users, roles, integrations |
| Billing | Manage subscription, view invoices |
| Super Admin | All permissions, manage SSO, approval workflows |
Compliance
- SOC 2 Type II — control mapping provided (CC6.1, CC6.3, CC7.2, CC7.3, CC7.4, CC8.1)
- GDPR — edge processes data in your infrastructure, SaaS receives only aggregated metadata
- Data residency — edge data stays in your infrastructure, SaaS region configurable (US/EU)
- Audit log — immutable, exportable, indefinite retention (Enterprise)
- MFA — supported at login for all SaaS accounts
Supply Chain & Build Integrity
- SBOM — generated for every runtime and CLI image on release
- Image signing — release images are signed with provenance attestations
- Vulnerability gate — releases are blocked on HIGH or CRITICAL CVEs in container images
- Nightly scan — automated vulnerability scan runs nightly against published images
Operational Security Readiness
Security controls must be operable under incident pressure, not just documented.
- Runbooks: reject spike, SaaS disconnect, bad bundle rollback, budget exhaustion — Runbooks
- Operations hub: consolidated incident handling and reliability controls — Operations Hub
- SLO/alert baseline: pre-defined checks for
no_bundle_loaded, reject spikes, descriptor mismatches, SaaS reachability — SLO and Alerting
Responsible Disclosure
If you find a security vulnerability, please report it to security@fairvisor.com. We commit to:
- Acknowledge within 24 hours
- Provide an initial assessment within 72 hours
- No legal action against good-faith reporters
- Credit in our security advisories (if desired)