Automation First, Regrets Later
summary
A platform built and maintained by two engineers manages 25,000+ resources and 65,000+ lines of Python using an Infrastructure as Software approach with Palumi, async Python outputs and a custom executor; circular dependencies force modeling tricks, previews grow slow, stacks split into backplane and module stages, and an SSM pin propagates attributes between stacks. The concrete regrets are long execution times, tangled code vs. deployment complexity, and a multi-year refactor strategy of tiny safe moves to avoid accidental resource churn.
AI-written summary · about
chapters (5)
- 0:00Automation philosophy
- 4:06Automation first approach
- 12:59Circular dependencies in code
- 22:39Refactoring production
- 25:30Automation regrets
related talks
55mThe Troubles of Automating “all the Things”
22mAutomation Run Rampant
38mPlatform Engineering in the Age of AI Teammates
02mThe Problems We're Here to Solve
21mAutomation as a Core IT Principle
38mWhy Automating Everything Adds to Your Toil
38mAutomation Is Hard & We Are Doing It Wrong
54mInfrastructure as code (anti) Patterns