Backup, Disaster Recovery and Exit

​

How does PrimeVault prevent downtime, and what are your disaster-recovery SLAs?

​

Architecture and backups

• Multi-AZ AWS architecture with automated failover and cross-region redundancy
• Hourly cross-region backups for application and system data, such as transaction metadata, policies, and logs
• Immutable storage for audit logs to preserve the trail after recovery

​

Operations

• 24×7 on-call engineering with documented runbooks for failover, rollback, and data restore
• Continuous health checks on indexers, ingest pipelines, and event publishers

​

Targets under disaster conditions

• Recovery Time Objective (RTO) under 4 hours for service restoration
• Recovery Point Objective (RPO) of 1 hour or better due to frequent backup cycles

​

If PrimeVault is unavailable, how do we stay independent and recover funds?

​

Key material protection

• MPC key shares are encrypted at rest with AES-256 and protected by hardware-backed KMS
• Only attested enclaves can unwrap shares during normal operations
• Encrypted share backups are stored in immutable, cross-region object storage
• You can also keep your own encrypted copies or use a third-party escrow

​

Your recovery path

• You retain encrypted MPC shares and an offline recovery kit
• With the required quorum, you can reconstruct vault keys locally and sweep funds without PrimeVault
• The design scales to tens of thousands of keys and supports cold, warm, and hot configurations
• Practiced teams typically regain full access in about 2 hours, assuming quorum participants are available

​

Recommended customer practice

• Mirror daily encrypted share backups to at least two regions you control
• Test reconstruction quarterly to validate keys, procedures, and quorum readiness