Building Resilient User Experiences

Large-scale sites are complex systems, where every page is likely generated with data pulled from multiple back-end data stores and services. Yet, not every piece of data on the page is critical to the user experience. The front-end of your site (i.e. front-end applications, templates, and user interface) needs to be designed to tolerate service outages. Features that your users will see need to have not just manual configuration flags, but also the ability to fail gracefully when a back-end service is unavailable. It’s important to note that in many cases, your site’s users will never realize the difference.

This talk is focused on getting front-end engineers, application developers, and product managers to think about their sites and products with an operational mindset. Topics covered in this talk will include:

• Identifying data sources used throughout your site
• Front-end design patterns for failure modes
• De-coupling non-critical user interfaces
• Effectively communicating failure to your community
• Validating failure scenarios with “Game Days”
• Configuration flags