Something About Distributed Systems

In this track you’ll hear from engineers and architects who are living with distributed systems every day—shipping features, firefighting incidents, and pushing the limits of scale. Talks will focus on:

• Latency in the wild – How teams measure, understand, and reduce user-perceived latency, deal with long tails, and design APIs, retries, and backpressure that behave well under load and partial failure.
• Consistency trade-offs – Concrete stories of choosing (and sometimes regretting) consistency models, handling stale reads and write conflicts, and designing systems that remain understandable as they grow.
• Failure as a first-class concern – Postmortems, incident narratives, chaos experiments, and the operational practices that make failures survivable instead of catastrophic.
• Scaling beyond “it works on my cluster” – Techniques and patterns for evolving architectures under growth: sharding, multi-region, multi-tenant, and cost-aware scaling strategies.
• Tooling and observability – The metrics, tracing, logging, and testing approaches that make complex distributed behaviors visible and debuggable.

Expect candid war stories, design explanations grounded in trade-offs, and patterns you can take back to your own systems. The emphasis is not on idealized architectures, but on the pragmatic decisions and hard-won lessons that separate distributed systems that merely run from those you can trust in production.