Infinite Scaling, Finite Failures: Serverless Resiliency Patterns and Lessons Learned

Katy Shimizu, Microsoft

You’ve built a beautiful workflow from serverless components. Your costs are down, your productivity is up, you’re scaling smoothly and life is good! ..now something is down, events keep coming in, and failures are cascading through your system. How do you recover from this? Can you?

The distributed nature of most serverless architectures makes resiliency, consistency and state management key design considerations. In this talk, we’ll discuss strategies to minimize the impact of platform and application failures and examine Azure’s Durable Functions as an example approach for creating resilient, stateful serverless processes and workflows.

Objective of the Talk

At the end of this talk the audience should be aware of some of the many ways a serverless application can (read: will) fail, and approaches to handle these eventualities.

Required audience experience

Familiarity with Azure and Azure Functions is helpful, but not required.

 

Track 3
Location: Burton & Redgrave Date: November 6, 2019 Time: 2:55 pm - 3:40 pm Katy Shimizu, Microsoft Katy Shimizu, Microsoft