Upgrading long-running, mission-critical workflows in production is one of the hardest challenges in distributed systems, especially when those workflows move money.

In this session, Daniel Skarbek, Senior Staff Software Engineer from Block, shares how the team has designed and upgraded core payment ledger workflows using Temporal, supporting a system that processes millions of transactions daily with strict reliability and correctness requirements. You’ll learn how Block implemented a dispatcher pattern for workflow versioning and how this approach enables safe upgrades even while workflows are actively running.

This webinar dives into the technical tradeoffs, upgrade strategies, and common pitfalls teams face when evolving long-lived workflows in production, and shows how Temporal’s durability and visibility make it possible to coordinate complex payment flows with confidence.

Key takeaways:
- Implement a dispatcher pattern for workflow versioning to safely route traffic between old and new workflow versions.
- Upgrade long-running workflows in production without breaking in-flight executions or compromising data integrity.
- Understand when built-in workflow versioning tools are enough, and when custom approaches are required.
- Avoid common pitfalls in workflow upgrades, including schema changes, behavior drift, and replay-related surprises.

---

Temporal is the simple, scalable, open source way to write and run reliable cloud applications. 

Learn more
Blog: https://temporal.io/blog
How Temporal Works: https://temporal.io/how-temporal-works
Community Slack: https://temporal.io/slack

Developer resources
Docs: https://docs.temporal.io
Courses: https://learn.temporal.io/courses
Support forum: https://community.temporal.io

Temporal's platform is a central hub for developers interested in distributed systems, workflow orchestration, and fault-tolerant computing. Through articles, tutorials, and documentation, Temporal offers insights into building resilient and scalable distributed applications. Developers can learn about workflow automation, state management, and handling failure scenarios using Temporal's workflow engine and coordination APIs.

Temporal

A Block (Square) engineer shares how they safely upgrade long-running payment processing workflows built on Temporal at scale (15M+ workflows/day). The talk covers three versioning strategies: worker versioning, patching, and their custom 'versioning by dispatch' approach. The dispatch pattern uses a single deployed app with a dispatcher class that routes workflows to the correct versioned implementation based on a LaunchDarkly feature flag stored in workflow input. To upgrade long-running workflows stuck in unbounded waits, they use Temporal's 'continue as new' feature triggered via a batch signal operation, migrating ~900K workflows at ~50/sec. Key lessons include always forking a new version when touching workflow code, using protobuf objects for forward/backward-compatible inputs, the 'fork behind' git pattern to preserve history, and always calling 'every handler finished' before continue-as-new.

Road to Replay: Safely Upgrading Payment Workflows at Scale with Block