Scaling the Stage: Managing 200+ Plugins & Thousands of Engineers
summary
Spotify describes running 200+ plugins in one ecosystem supporting thousands of engineers and compares three concrete scaling choices: replicate the full Backstage instance behind a load balancer, split heavy plugins into separate instances with conditional enabling, or deploy plugins separately as one-repo, one-backend apps. Key implementation details: plugins discover each other via a configurable Discovery service (or static app config), secret injection via cloud service accounts, and repo orchestration with tools like Fleet shift to manage many deployments.
AI-written summary · about
chapters (9)
- 0:00Backstage backend scaling
- 1:25Plugin communication
- 5:42Discovery service & deployment
- 8:14Replicating Backstage instances
- 10:00Resource-heavy plugins
- 11:25Independent plugin deployment
- 12:50Spotify’s Backstage migration
- 15:39Scaling Backstage advice
- 17:58Q&A
related talks
19mHigh Performance Backstage for High Performing Organizations
26mBackstage: Shaping the Future Of Developer Experience
06mYour Backstage, Our Grunt Work: Migrating to Sp...
27mShifting Spotify Engineering from Spreadsheets to Backstage
23mScaling the Sound: Fleet Management at Spotify
19mBackstage Plugins: Another way to communicate your data
12mValue Driven Platform Engineering Using Backstage
19mBackstage Plugins: Another way to communicate your data