Why You’re (Probably) Doing Service Catalogs Wrong
summary
Service catalogs often take months or years to ship and then languish because teams don’t have the data or the incentive to keep it fresh—efforts spun up a paid tool, crawled Kubernetes pods, spent weeks mapping cryptic pod names to services, and still saw low adoption and rotten data. Practical fixes: start with one team and deliver a working artifact in weeks, make catalog actionable (pager routing, deploy-to-cluster mapping), and solve the chicken-and-egg data problem by iterating with champions and editable source data.
AI-written summary · about
chapters (9)
- 0:00Service catalogs intro
- 2:01What a service catalog should be
- 3:02Struggles building catalogs
- 10:03Three big problems
- 11:07Problem: Unclear objectives
- 12:11Problem: Lack of buy-in
- 15:00Problem: Data decay
- 17:29Key takeaways
- 18:12Q&A
related talks
25mWhy you’re doing service catalogs wrong
31mRethinking Data Catalogs: The Promise and Pitfalls
52mScaling Data Teams: Three Pitfalls and How To Avoid Them
32mSIG Service Catalog Deep-Dive
23m3-Step Path for Fixing Data Discoverability
24mThe Lifecycle of A Service
30mWorking with DBAs in a DevOps World
32mTales from the DevOps Transformation Trenches