Service Oriented Disasters

on YouTube ↗26m

" summary

Three concrete case studies: extracting a votable hats service fails to be independent because it connects to the old DB and reuses tangled hat models, packaged as a gem and backed out before production; an identity service without its own database that creates a hard coupling and a single point of failure so both apps must stay up and scaling worsens (example numbers: 2,000 main servers vs 100 auth servers); and a successful pattern: dark‑launching rewrites with GitHub’s Scientist to compare old and new paths in production.

AI-written summary · about

" chapters (6)

1. 0:00Start
2. 0:41bit.ly/ service oriented disasters
3. 2:321. Make each class and method conceptually simple.
4. 9:09Problems to solve: 1. Follow Single Respon
5. 15:17Case Study 2: An Identity Service
6. 20:56Case Study 3: Sharing Assets Across Application

" related talks

• 30mService Oriented Design in PracticeBryan Helmkamp · MountainWest RubyConf 2011
• 01h 00mService Oriented Architecture for Robust & Scalable SystemsOle Michaelis · PHP UK Conference 2014
• 39mStop Building Services, Episode 1: The Phantom MenaceRachel Myers · ArrrrCamp 2015
• 54mServices, Services, Everywhere!Evan Phoenix · Barcelona Ruby Conference 2014
• 36mAn Iterative Approach to Service Oriented ArchitectureEric Saxby · RailsConf 2014
• 35mMicroservices: A Tale of 3 HeroesSha Ma, Jovel Crisostomo · Craft Conference 2021
• 39m10 Tips for failing at microservicesDavid Schmitz · code.talks 2017
• 41mMe, My Organization and My-croservicesKrzysztof Mejka · JDD 2017