Active-Active Multi-Region Microservices

The problem

The existing setup ran active-passive across regions, meaning a full regional outage during peak seasonal trading (the highest-risk time to be down) relied on a manual or semi-automated failover with meaningful recovery time.

What changed

• Re-architected the microservices to run active-active across multiple Azure regions on AKS, with traffic routed to whichever region is healthy rather than failing over from a single passive standby.
• Built out data replication and conflict-handling strategies so both regions can serve live traffic without stepping on each other.
• Ran structured DR and failover game days, measuring recovery time objective (RTO) before and after each round of changes.

Result

• 100% service continuity maintained during peak seasonal trading, even through simulated regional failures.
• Potential RTO cut by 40% versus the previous active-passive design.
• Failover is now largely automatic rather than depending on manual intervention under pressure.