Presentation: Scaling Slack - The Good, the Unexpected, and the Road Ahead

Slack is a persistent communication app for teams, with high customer expectations to deliver a reliable, rich, low latency client experience. 

Over the past couple of years, we've made major changes to the core service architecture to meet these needs for larger and larger enterprise customers. This talk will present three of these changes -- decomposition of the real-time message service, client-side lazy loading via edge caching, and scaling the primary data storage tier with fine-grained horizontal sharding using Vitess.

Each of these projects has resulted in significant improvements to the reliability and performance of the Slack service, yet ran into some bumps along the way during the deployment process. Still, we're far from done with this journey, so we'll touch on some of the pain points that remain and high-level thoughts for how we plan to address them.