TopolearnA shared framework for trade-offs keeps designs reviewable, and DR strategy determines what an outage costs you.
The AWS Well-Architected Framework organises design questions into six pillars:
- Operational excellence: infrastructure as code, small reversible changes, observability, runbooks.
- Security: strong identity, least privilege, traceability, encryption, protection at every layer.
- Reliability: recover automatically, scale horizontally, test recovery, manage change.
- Performance efficiency: the right resource types, serverless where it fits, measure and experiment.
- Cost optimization: pay only for what you need and attribute spend.
- Sustainability: maximise utilisation and minimise resources needed.
The free Well-Architected Tool walks a workload through these questions (plus lenses such as Serverless and SaaS) and tracks improvement items.
For resilience, start within a Region: spread across at least two AZs, keep compute stateless behind load balancers, use managed multi-AZ data stores, and design for static stability, meaning the system keeps working through an AZ loss with capacity already in place rather than needing to launch new resources mid-incident.
For disaster recovery, define two targets per workload: RPO (how much data you can lose) and RTO (how long you can be down). Then pick a strategy, in rising cost:
- Backup and restore: hours of RTO; backups copied to another Region.
- Pilot light: data replicated continuously, core infrastructure defined but scaled to zero.
- Warm standby: a scaled-down, fully working copy ready to scale up.
- Multi-site active/active: near-zero RTO, the most complex and expensive.
A DR plan you haven't tested is a hope, not a plan: rehearse restores and failovers regularly.
Example
{
"BackupPlanName": "prod-daily",
"Rules": [
{
"RuleName": "daily-35d-with-dr-copy",
"TargetBackupVaultName": "prod-vault",
"ScheduleExpression": "cron(0 3 * * ? *)",
"StartWindowMinutes": 60,
"Lifecycle": {"DeleteAfterDays": 35},
"CopyActions": [
{
"DestinationBackupVaultArn": "arn:aws:backup:eu-central-1:111122223333:backup-vault:dr-vault",
"Lifecycle": {"DeleteAfterDays": 35}
}
]
}
]
}Remember
- Six pillars: operations, security, reliability, performance, cost, sustainability
- Multi-AZ and static stability first; multi-Region for disaster recovery
- RPO = acceptable data loss, RTO = acceptable downtime; they pick the DR strategy
- Test recovery regularly
Watch out for
- Treating backups that have never been restored as a disaster-recovery plan.