The CTO Podcast with Fexingo · 2026-07-04 · 8 min
How Uber rebuilt its dispatch engine to handle 30 million daily rides without a single downtime event. Lucas and Luna break down the core architectural shift from a monolithic matching service to a fully distributed, event-driven system that uses probabilistic matching and decentralized state. They discuss the surprising failure mode during the rollout - a subtle race condition in the geohash partitioning logic that caused phantom driver surges - and how Uber's team fixed it by introducing a two-phase commit with a lightweight consensus layer. Along the way, they explore trade-offs between latency and consistency, the role of observability in catching rare bugs, and why Uber chose to keep its dispatch logic in-house rather than buying a third-party solution. A focused look at how one of the world's largest real-time systems evolves under the pressure of scale.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.
Durable Execution for Real‑World Failures with Temporal’s Cornelia DavisPlatform Engineering Podcast · on Event-driven architecture96 / 100
TigerBeetle's Spectacular Jepsen Report - with Joran GreefSoftware Unscripted · on Distributed Systems88 / 100
TCG079: Why Your State File is Actually a Distributed Systems ProblemThe Everything Feed · on Distributed Systems76 / 100
Tech interviews with NeetCodeThe Pragmatic Engineer · on Distributed Systems74 / 100
Embedding Value in the Insurance MarketTech means Business · on Event-driven architecture63 / 100
What does code mean in 2026?Thoughtworks Technology Podcast · on Distributed Systems61 / 100
How Airbnb Rebuilt Its Search for 100 Million Listings72 / 100How GitHub Migrated 100 Million Repositories to a New Storage Engine76 / 100How Dropbox Rebuilt Its Sync Engine for 700 Million Users85 / 100How Skyscanner Migrated 300 Microservices to Event-Driven Architecture85 / 100How Palantir Rebuilt Its Foundry Ontology for Government AI Deployments85 / 100