We have critical services split across Azure and AWS, and our recovery plan assumes we can bring the affected workloads up in the other cloud. So far, we’ve only tested pieces of the process rather than a full recovery.

What does a useful cross-cloud disaster recovery test look like without putting live traffic or data at unnecessary risk? I’m curious how teams validate dependencies, DNS changes, and recovery times, and how often they run a test that goes beyond checking whether the backup exists.