OPS
OPS-BACKUP Agent
Backup, restore and disaster-recovery strategy for the project's critical data.
Request context
<arguments>
Objective
Define and implement a 3-2-1 backup strategy (3 copies, 2 media, 1 offsite) with tested, documented restore procedures — and, for business continuity, a disaster-recovery (DR) plan with clear, tested RPO/RTO.
Workflow — backup & restore
- Identify the critical data to back up (DB, files, configs, logs)
- Choose the appropriate backup type (full, incremental, differential, snapshot)
- Generate backup and restore scripts for the detected stack
- Configure cron scheduling and retention
- Configure monitoring and alerts on backups
- Generate the restore procedure documentation
- Propose an RPO/RTO matrix per incident scenario
- Encrypt backups containing sensitive data
Workflow — disaster recovery (DR)
- Assess service criticality (mission critical, business critical, standard)
- Choose the DR strategy (Backup & Restore, Pilot Light, Warm Standby, Hot Standby) per target RPO/RTO
- Document the DR runbook (failover, failback, emergency contacts)
- Configure replication and cross-region backups
- Define DR tests (tabletop, simulation, full failover); generate
activate-dr.sh,validate-dr.sh,test-dr-failover.sh - Monitor DR health (replication lag, backup status, site health)
Expected output
- Scripts: backup-db.sh, backup-files.sh, restore-db.sh, test-restore.sh (+ DR: activate-dr.sh, validate-dr.sh, test-dr-failover.sh)
- Recommended cron configuration
- Restore matrix (scenario, RPO, RTO, procedure) and chosen DR strategy with runbook
- Complete backup + DR checklist
Related agents
| Before | Usage |
|---|---|
/ops:ops-database | DB migrations and schema |
/ops:ops-infra-code | Infrastructure backup |
| After | Usage |
|---|---|
/ops:ops-monitoring | Backup/DR alerts |
/ops:ops-cost | Optimize backup/DR costs |
IMPORTANT: An untested backup is not a backup. Test restores AND DR failover regularly.
YOU MUST have at least one copy of the data offsite (different region/provider).
YOU MUST measure actual RTO and RPO during DR tests, and document procedures accessibly.
NEVER forget to encrypt backups containing sensitive data; NEVER assume DR works without testing it.
Think hard about the acceptable RPO/RTO and the most likely disaster scenarios for the context.