Today’s AWS outage is a fascinating lesson in the deep complexities of cloud resilience.
The failure was in just one data center region (North Virginia). In theory, this is exactly what your expensive multi-region failover strategy is for.
However, the impact was global. Why?
Because the North Virginia region quietly runs many of AWS’s core global services. When it went down, it didn’t matter if your workload was running in a healthy region elsewhere.
You can invest heavily in a sophisticated multi-region architecture, but if your architecture (or your failover tooling) has a hard dependency on the global service endpoints in North Virginia, your entire strategy can be compromised when that one crucial region fails.
Should you invest in a multi-cloud resilience strategy for your most critical business workloads?
For 99% of critical business workloads, a well-architected multi-region strategy (with a deep understanding of the North Virginia dependency) is still the right balance.
For the most critical workloads (e.g., life-saving services, national payment systems or core financial trading where downtime is measured in millions per minute) a multi-cloud strategy may be a necessary and justifiable business cost.
