Apiumhub had the pleasure to interview Rene Weiß – the Global Software Architecture Summit speaker, to get to know what are the key Software Architecture insights he could share with GSAS attendees and learn more about his chapter in the Software Architecture Metrics book that was recently published by O’Reilly.
About Rene Weiß
Rene Weiß, CTO at fintech FINABRO and independent agile and software architecture consultant.
What Software Architecture Metrics do you normally use?
I like a combination of a few easy and basic ones combined with some more complicated metrics giving a higher level of abstraction and feedback:
- Basic Code Metrics:
- Cyclomatic complexity on the code level,
- Unit test coverage,
- Integration test coverage to give a brief overview of the basic “fitness” of the code basis;
=> usually I am more interested in the overall development of those 3 numbers than the distinctive single value at a specific time
- Performance measures for key parts of the solution (“do we get slower” in key areas) – I would measure in a test system as part of a CI / Test build
- Response times for key parts of the solution (either web service response times or page load times using google lighthouse) – these should be monitored in the live system
- High-level metrics: 1-3 business metrics showing the healthiness of the solution in business terms (e.g. average sales per day (or per week) that would show deviations/problems on the very high level
- + Some team metrics (not directly related to architecture)
- A rolling average (e.g. 2 weeks) of new bugs/incidents vs. Bugs/incidents closed
- In larger endeavors the deployment frequency (= number of deployments/timeframe)
What are the key Software Architecture insights you could share with GSAS attendees?
- Evolutionary architectures and fitness function: What are they and how to use them
- How to use the fitness function concept to define metrics and how that can help enable the whole team to participate in software architecture work
Can you briefly comment on your software architecture metrics book chapter?
I introduce the idea of the fitness function testing pyramid which is closely related to the testing pyramid that most of us have heard of.
Using this metaphor, the fitness function testing pyramid enables us to balance the right amount of architecture tests that produce metrics we can use to enhance and build the solution by still having some safety net.
If you liked this interview you can subscribe to our Youtube channel. We will be posting there interviews with other speakers, testimonials, and talks from the event.