Cloud resilience isn’t just about keeping the lights on — it’s about designing systems that expect things to go wrong and still deliver. In this session, we explore what it really means to build for failure in Azure. We’ll start with the platform layer: what happens under the hood with Zonal Redundancy (ZRS) and Geographic Redundancy (GRS), how the Azure platform itself is evolving to support higher resilience, and what you can (and should) lean on when designing robust infrastructure. Then we zoom into Analytics as one of the many mission critical workloads customers run on Azure — because when your data stops flowing, so does your business. From ingestion pipelines to distributed storage and query engines, we’ll walk through best practices and real-world strategies to make your data platforms more fault-tolerant and redundant without breaking the bank (or your SLAs). Whether you're architecting core platforms or scaling enterprise analytics, this talk is your resilience cheat code — practical, opinionated, and failure-friendly. Key Takeaways Understand the Mechanics of Azure Resilience Learn how Zonal Redundancy (ZRS) and Geographic Redundancy (GRS) work behind the scenes — and how to apply them effectively. Design for Failure, Not Just Uptime Shift your mindset from “avoid failure” to “absorb and recover” using platform-native tools, design patterns, and deployment strategies. Bridge the Gap Between Platform and Workload Resilience See how platform-level redundancy sets the foundation — but workload-level design completes the story, especially for analytics systems. Apply Real-World Resilience Patterns to Analytics Workloads Discover concrete guidance to keep data pipelines, data lakes, and query layers highly available and redundant across zones and regions. Avoid Common Pitfalls and Overengineering Traps Not all services are equally resilient — learn when to build, when to trust the platform, and when you're paying too much for too little. Prepare for Disasters Before They Happen Equip yourself with actionable strategies for disaster recovery, incident response, and architecture reviews that matter. NOTE: Although this session doesn't cover Observability in depth, combining it with Tom Claes' session titled "Observe. Predict. Prevent. - Powered by Azure Monitor" would create a great 2-part series to provide attendees a complete view of resilience in Cloud.