Distributed Locks, Sagas & Coordination with Cosmos DB (Azure Cosmos DB Conf 2026)

Microsoft Developer hosts a session from Azure Cosmos DB Conf 2026 where Eric Boyd explains how to use Azure Cosmos DB as a coordination mechanism in distributed systems—especially when retries, concurrency, and multi-region deployments make “just put a lock around it” fail in production.

Full summary based on transcript

What goes wrong with “just add a lock” in distributed systems

Eric Boyd frames the problem as coordination under real production conditions:

Using Azure Cosmos DB for coordination

The session focuses on using Cosmos DB as a coordination mechanism for distributed workflows, including:

Correctness tradeoffs to plan for

Eric Boyd emphasizes correctness and safety tradeoffs teams hit in production, including:

Multi-region considerations

The talk highlights that multi-region deployments change the coordination problem:

Presenter