Disaster Recovery

Blockchain data is immutable and doesn't need backing up — deployed bytecode, transaction history, and on-chain state are permanent by definition. What does need protecting is everything the platform manages around your contracts: configuration, metadata, version history, access control, monitoring rules, operational state, and audit logs.

All of it is backed up automatically, verified for integrity, and recoverable to any point within your retention window.

What gets backed up

Configuration and metadata — contract records, version history with changelogs, deployment records and transaction hashes, access control configurations, alert rules and monitoring settings, and API credentials stored encrypted.

Operational data — execution history, health metrics, integration configurations, and webhook delivery records.

Audit logs — the complete, append-only record of every platform operation.

On-chain data is not backed up because it doesn't need to be. If you need to verify on-chain state, you query the chain directly.

Automated backup schedule

Backups run on a predictable schedule without any manual intervention:

  • Full backup — runs weekly, captures everything
  • Incremental backup — runs daily, captures changes since the last backup

Both types include encryption, compression, and automatic checksum verification after completion. If checksum verification fails, the backup is marked corrupted and an alert fires immediately — you know about a bad backup right away, not when you try to use it.

Schedules and retention windows are configurable.

Recovery tiers

Different contracts have different recovery needs. The platform offers three tiers:

| Tier | Recovery Point Objective | Recovery Time Objective | Backup Frequency | |------|--------------------------|------------------------|------------------| | Standard | 24 hours | 4 hours | Daily | | Enhanced | 4 hours | 1 hour | Every 6 hours | | Critical | 1 hour | 30 minutes | Continuous replication |

Recovery tiers are configured per contract. High-value production contracts can run on Critical tier while development contracts run Standard.

Point-in-time restore

Restore your platform state to any point within your retention window — not just to the last backup. Select the target timestamp, select the target environment (production, staging, or development), and the platform reconstructs the state as it existed at that moment.

Before committing to a restore, run a dry-run pass first. The dry run shows exactly what would be restored, what would change, and what the resulting state would look like — without actually making any changes. This is not optional: always run the dry run first.

Selective restore

If an issue affects only a subset of contracts, there's no need to restore everything. Selective restore targets specific contract IDs, restoring only those records while leaving everything else unchanged.

Quarterly recovery testing

Recovery plans that aren't tested are plans that fail at the worst time. Quarterly restore tests run automatically: they restore a recent backup to an isolated environment and verify the result — checksum integrity, record counts, version history completeness, and access control preservation.

Results are logged and available for compliance audits. You always know that your backup is recoverable, not just that the backup completed.

Manual backups

Trigger a manual full backup before any high-risk operation: a major version upgrade, a migration, an access control overhaul. The backup is labeled with your description and excluded from automatic cleanup by default — it persists until you explicitly delete it.

Backup retention

Retention windows are configurable:

  • Full backups: 90 days by default
  • Incremental backups: 30 days by default
  • Manual backups: retained indefinitely by default

Old backups are cleaned up automatically per your retention policy.

Alerts

Backup failures, missed backup windows, and recovery time objective violations all fire alerts to configured notification recipients. If a backup fails, you know immediately — not when you're in the middle of a recovery operation.


Contract lifecycle → Multi-chain support → Security →