SUGCON 2024 Europe | Bulding for Resilliency by Andy Cohen

 

Andy Cohen captivated the audience this afternoon at SUGCON Europe 2024 with his session on "Building for Resiliency". I was still determining what to expect from the session, given its broad title, but I've valued Andy's insights from our past conversations. And as expected, he delivered valuable takeaways that left me contemplating after the session ended.

The session delved into the architecture of XM Cloud, offering a case study on resilience. Andy walked us through the journey of XM Cloud from ideation to its current state, highlighting pivotal moments and lessons learned along the way.

The presentation kicked off with an overview of how the idea for XM Cloud originated within Sitecore, providing insights into the early stages of development and the dedicated team behind it. While this part may have felt somewhat ordinary to some, it laid the groundwork for the more interesting topic to come.

As Andy delved deeper into the technical aspects, the session came to life. He tackled the challenge of building a scalable SaaS application with external dependencies head-on. One standout example was their approach to implementing health checks, a fundamental aspect of any resilient system. Andy shared freely about their initial failures in this regard, particularly their reliance on GitHub's health status as an indicator for system health. Through trial and error, they discovered the importance of granularity in health checks, enabling them to respond effectively to degraded services without compromising the entire system.

The session concluded with a great insights distilled from their XM Cloud's journey:

• Embrace adaptability and flexibility in your codebase.
• Leverage microservice architecture to segment domains effectively, promoting flexibility.
• Implement continuous delivery practices with feature toggles to enable seamless updates.
• Foster a culture of sharing common patterns and libraries to streamline development.
• Utilize tools like Kafka for efficient event-driven architectures.
• Prioritize resilience in authentication, background workers, and data access layers.
• Release common libraries with a versioning strategy to avoid breaking changes.
• Maintain microservices at a manageable size to simplify maintenance.
• Implement retry mechanisms in your services, leveraging tools like Polly.
• Continuously educate yourself on resilience through books and blog posts.

Overall, Andy's session emphasized the importance of resilience in modern software architecture and provided actionable insights for building robust systems. It served as a reminder that in an ever-evolving technological landscape, granual adaptability and resilience are paramount.