Modern Platform Engineering: 9 Secrets of Generative Teams • Liz Fong-Jones • GOTO 2023
Interesting. report from Liz Fong-Jones, Field CTO from Honeycomb. Liza first remembers what DevOps is, then SRE, and then moves on to Platform Engineering. In the part about devops we are talking about the desire for culture, automation, lean, measurement and sharing. Dora metrics measure how well teams do. But these metrics have a number of problems:
- The path of improvement is not too clear.
- The improvements themselves are hard to measure. (Plus, we're measuring lag rates.) There is a dilemma with the maturity model itself and continuous improvement. (Maturity models assume that you can get to a certain level of maturity, and then you do not need to improve - by the way, Accerate proposed using capabilities model to solve this problem. told about it in the book review) How to achieve consistency in different teams
As a result, Liz offers a model with 9 secrets that contribute to the generative culture of Westrum (I already am. wrote This approach to the typification of cultures) 1. Reproducible deploys (If it ain't in Git it don't exist!) - We need repeatability in assemblies and deployments. This was discussed at the time of appearance. 12 factor app, and now it is a de facto standard:) 2. Fast CI/CD (I wanna go fast!) We speed up feedback cycles and help developers see the results of their work faster. 3. Observability (Don't drive with snow on your windshield!) - default to understand what is happening to our application. Without this, it is not clear how to identify problems, as well as thrash them. 4. Feature flagging (Smaller boom, less fear!) - complete with TBD (trunk-based development) This allows you to do CI correctly. (more report) 5. Code ownership (You build it, you run it! (with a twist)) Mantra that removes the fence between dev and ops (sre) teams. In fact, SDEs (software development engineers) own not only the code, but also the application itself, which revolves in production 6. Blameless culture (No one wins the shame game!) Here I will recommend several articles and speeches.
- Google's Project Aristotle What makes a team effective at Google?
- A typology of organisational cultures An interesting study on the typology of organizational cultures (pathological, bureaucratic, generative) My presentations on related topics
- Post-mortem culture
- From monolith to microservices and back 7. Service level objectives (SLOs) (If you liked it you shoulda put an SLO on it!) - About me. spoke “Designing Reliable Systems – Is the Game Worth the Candle” 8. Chaos Engineering (Test your assumptions!) I recommend you to look at the presentation on this topic. Kelly Shortridge "Practical Magic: The Resilience Potion & Security Chaos Engineering" 9. Platform teams (to tie it all together and systematise it) - I recommend reading the book. Team Topologies About team-first approach in designing architecture of software systems and organization. It speaks very well about platform teams. Here’s how Liz describes these teams.
Platform teams outsource as much as possible, write as little code as possible. Platform teams are experts in outsourcing. It’s a very high-leverage role; they use their infra expertise to offload as much operational burden as possible. As a result, in the Liz model, platform teams help everyone else. 8 They allow you to build a generative culture and highly effective teams.
#Engineering #SystemEngineering #SystemDesign #SoftwareArchitecture #Software #Devops #PlatformEngineering #Process #Culture #SRE