Atlassian successfully migrated 4 million Jira databases from AWS RDS PostgreSQL to Aurora PostgreSQL across 3,000 server instances in 13 regions. The migration faced a critical challenge when Aurora's file count limitations prevented startup with 20+ million files per cluster. They solved this by developing a "drain-convert-refill" strategy that reduced tenant density during conversion, then repopulated clusters afterward. The project achieved 99.99% uptime SLA, better cost efficiency through auto-scaling, and maintained under 3 minutes downtime per tenant while processing up to 90,000 daily migrations at peak.

12m read timeFrom blog.bytebytego.com
Post cover image
Table of contents
✂️ Cut your QA cycles down to minutes with QA Wolf (Sponsored)Migration StrategyThe File Count Limit ProblemThe Draining SolutionConclusionByteByteGo Technical Interview Prep KitSPONSOR US

Sort: