Platform Health: A team model of platform evolution
summary
A small platform team at Jellyfish balances scalability, reliability, developer velocity and FinOps while the company grows from under 20 to over 200 people and the engineering org to ~50. The team actually replaced our operational data store then pivots under load, adopting a more familiar DB hack as a pivot to a Band-Aid solution that stabilizes production. Work style is DIY product management, heavy on allocations and sunk‑cost dollar metrics, staffed with platform generalists using Python Bash Terraform and SQL.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 2:23Crisis and team challenges
- 5:33Platform health and team scope
- 10:01Paved road and platform team
- 12:35Tenets for platform decisions
- 16:54Eisenhower Matrix and cognitive load
- 19:22Team identity and MVP
- 21:08Team composition and skills
- 23:09Non-obvious metrics
- 25:58Database pivot and resilience
- 28:20Q&A
related talks
47mPlatform Team Structures and Concerns
25mBreaking the 2 Pizza Paradox with Platform Applications
32mHow we did Platform Engineering at Datadog
15mPlatform Engineering - About Tools...
39mPlatform Engineering as a (Community) Service
29mIaC Next Steps in Platform Engineering
33mPlatform as a Product - Transforming from Service Delivery into Continuous Operations
42mPlatform Engineering in practice — Why and How to start