Case Study: Duolingo cuts a 500-directory Terraform apply from 5 days to a few hours with Spacelift
Key results
The challenge
Before adopting Spacelift in October 2024, Duolingo managed Terraform with Atlantis but found it hard to keep stable at scale. Constraints included an inability to scale past four concurrent plans and one concurrent apply, limited merge-before-apply support, and a slow PR-lock workflow.
The solution
Duolingo moved to Spacelift, tuning worker instance sizes and using Spacelift's hooks to customize its particular Terraform workflows without altering the core Terraform.
“Spacelift is robust, reliable, and customizable. There's a lot less babysitting and worrying; it's a lot smoother.”
ESEvan StratSenior Platform Engineer, Duolingo
The results, in context
Applying 500 directories, which previously took about five days in Atlantis, now takes a few hours in Spacelift. The team reports far less operational babysitting, freeing engineers to focus on higher-value problems.