Скачать или смотреть Serverless Job Execution at Scale: A Migration Story from Spring Boot Cron Jobs | Luke Hackett

Legacy applications that rely on cron jobs often become bottlenecks as organisations scale. In this talk, we’ll share our journey migrating a monolithic, cron-based job execution application built with Spring Boot into a modern, cloud-native, event-driven architecture capable of scaling to thousands of jobs per minute.

Our legacy application used cron expressions to schedule jobs that generate several hundred rails of content. While the application was deployed across multiple nodes, a key limitation in the design meant that only one node could execute these jobs. As traffic and workloads increased, this architecture quickly became a bottleneck—leading to missed job executions and forcing us to vertically scale all nodes, driving up costs and complexity.

The inability to horizontally scale scheduled jobs made it clear: we needed a more scalable and resilient solution.

We’ll share how we:
Replaced traditional cron jobs with Amazon EventBridge rules for distributed, event-driven execution
Used Amazon SQS and AWS Lambda for decoupling and fan-out processing, enabling parallel execution at scale
Refactored our job logic into modular, Lambda-compatible Java functions
Tackled common issues like message deduplication, database connection throttling and observing the stability of our system

Additionally, we’ll cover the nuances of migrating a production system without downtime, how we preserved job semantics like retries and idempotency, and how we built a framework to handle complex workflows and dependencies.

This session is perfect for developers and architects who are looking for an alternative way of processing jobs with minimal fuss. Attendees will walk away with some practical tips and lessons learned from transforming a rigid job execution model into a flexible, resilient, cloud-native system.