SFScon 21 - Eduardo Guerra - A Lean Software Analytics Canvas for Agile Small Teams
0 likes144 views
The document discusses a lean software analytics canvas designed for agile small teams to help make meaningful decisions based on metrics. It emphasizes the importance of using actionable data to address project issues, improve development processes, and optimize performance. Key recommendations include implementing monitoring tools, asynchronous report processing, and creating a report queue to manage system overload.
SFScon 21 - Eduardo Guerra - A Lean Software Analytics Canvas for Agile Small Teams
- 1. A Lean Software Analytics Canvas for Agile Small Teams Actually using metrics for meaningful decisions Eduardo Guerra SFSCon 2021
- 3. I think the clients would like more the user interface if we change it. I think we will have more clients and we need another server.
- 4. I think this algorithm can be a problem for performance and we need to optimize it. I think this part of the code needs urgent refactoring.
- 5. How development team make decisions Based on our experience!
- 6. Based on our experience! The experience How development team make decisions
- 7. I THINK For the resources that you are going to spend doing that, you need to be sure!
- 12. But what is the question?
- 13. Finding what metrics can help to solve the project problems. Figuring out what I can do with the metrics from tools that I have installed.
- 14. meet analytics
- 15. Analytics “Analytics is a data-driven approach to support decision making widely used in marketing in order to reach and better understand customers.” (Danvenport, 2009)
- 16. Software Analytics “Software Analytics is to enable software practitioners to perform data exploration and analysis in order to obtain insightful and actionable information for data-driven tasks around software and services.” (Zhang et al., 2011)
- 18. Development practices Testing methods Bug treatment Teams and collaboration Productivity Reuse of components Customer and requirements Maintenance and software evolution Evaluating quality Software usage Services
- 19. Development Tools Bug Reports Application Data Code Repository Running Software Source code Documentation
- 20. Every measurement should be done for a reason. The entire team need to be aware of the analytics process. The project cannot be stopped just to implement analytics. The result should be decisions and actions, and not numbers.
- 21. How to make that suitable for small and medium size teams?
- 23. goal metric decision Things are spread all over the development tools and documents!
- 27. Why are sometimes reports take too long to be generated by the software?
- 28. Why are sometimes reports take too long to be generated by the software? Request Rates Response Time CPU usage
- 29. Why are sometimes reports take too long to be generated by the software? Request Rates Response Time CPU usage (a) start and end time logs for report execution. (b) a tool to monitor CPU usage.
- 30. Why are sometimes reports take too long to be generated by the software? Request Rates Response Time CPU usage (a) start and end time logs for report execution. (b) a tool to monitor CPU usage. Configure tool to monitor the CPU on the server. Implement request logs into report feature. Analyze runtime logs.
- 31. Why are sometimes reports take too long to be generated by the software? Request Rates Response Time CPU usage (a) start and end time logs for report execution. (b) a tool to monitor CPU usage. Configure tool to monitor the CPU on the server. Implement request logs into report feature. Analyze runtime logs.
- 32. Why are sometimes reports take too long to be generated by the software? Request Rates Response Time CPU usage (a) start and end time logs for report execution. (b) a tool to monitor CPU usage. Configure tool to monitor the CPU on the server. Implement request logs into report feature. Analyze runtime logs. A large delay when multiple reports are requested at the same time.
- 33. Why are sometimes reports take too long to be generated by the software? Request Rates Response Time CPU usage (a) start and end time logs for report execution. (b) a tool to monitor CPU usage. Configure tool to monitor the CPU on the server. Implement request logs into report feature. Analyze runtime logs. A large delay when multiple reports are requested at the same time.
- 34. Why are sometimes reports take too long to be generated by the software? Request Rates Response Time CPU usage (a) start and end time logs for report execution. (b) a tool to monitor CPU usage. Configure tool to monitor the CPU on the server. Implement request logs into report feature. Analyze runtime logs. A large delay when multiple reports are requested at the same time. CPU usage is extremely high when reports are requested at the same time
- 35. Why are sometimes reports take too long to be generated by the software? Request Rates Response Time CPU usage (a) start and end time logs for report execution. (b) a tool to monitor CPU usage. Configure tool to monitor the CPU on the server. Implement request logs into report feature. Analyze runtime logs. A large delay when multiple reports are requested at the same time. CPU usage is extremely high when reports are requested at the same time When a report takes time to process, the user request it again, overloading the system. Reports does not need an immediate return.
- 36. Why are sometimes reports take too long to be generated by the software? Request Rates Response Time CPU usage (a) start and end time logs for report execution. (b) a tool to monitor CPU usage. Configure tool to monitor the CPU on the server. Implement request logs into report feature. Analyze runtime logs. A large delay when multiple reports are requested at the same time. CPU usage is extremely high when reports are requested at the same time When a report takes time to process, the user request it again, overloading the system. Reports does not need an immediate return. CPU usage should be lower than 90%. Make processing of reports asynchronous to avoid CPU overhead.
- 37. Why are sometimes reports take too long to be generated by the software? Request Rates Response Time CPU usage (a) start and end time logs for report execution. (b) a tool to monitor CPU usage. Configure tool to monitor the CPU on the server. Implement request logs into report feature. Analyze runtime logs. A large delay when multiple reports are requested at the same time. CPU usage is extremely high when reports are requested at the same time When a report takes time to process, the user request it again, overloading the system. Reports does not need an immediate return. CPU usage should be lower than 90%. Make processing of reports asynchronous to avoid CPU overhead. Implement report queue. Implement CPU alert on server.
- 38. The Lean Analytics Canvas helps to reason about the how the metrics can be used!
- 39. Software Startup Case studies introduced successfully the use of the canvas in development teams! Multinational Company