How do you plan disaster recovery?

Direct answer

What a strong answer covers

How it works

Disaster recovery starts with two numbers per service agreed with the business: the recovery time objective, how long it may be down, and the recovery point objective, how much data may be lost. Those numbers drive the choice between restoring from backups, keeping a pilot-light environment, a warm standby, or active-active across regions, with cost rising along that scale. The plan maps every dependency and data store, including the ones nobody thinks of such as DNS and secrets, automates the rebuild through infrastructure code, keeps credentials and runbooks accessible when the primary is gone, and is exercised on a schedule by actually failing over and running from the recovery environment.

Example

The billing platform has an RTO of 4 hours and an RPO of 5 minutes, so it runs a warm standby in a second region with streaming database replication and infrastructure defined in Terraform. Quarterly, the team fails over for real, serves traffic from the standby for an hour, and fails back; the first drill found that the secrets manager was single-region and the standby could not decrypt anything, which was fixed before it mattered.

Common mistakes

• Choosing a DR tier without a business-agreed RTO and RPO.
• Backing up the database but not the credentials and configuration needed to run it.
• Never running a failover so the plan exists only as a document.

A useful answer structure

Answer from your own work.

Do not memorise a script: use only examples you can explain honestly in follow-up questions.