Intro

L3 bug load is a concern in the SUSE Manager Development Team, and we want to do something about that.

Most developers have gut feelings (are they growing? what kinds are most frequent? how many are invalid?) - but we lack concrete insights.

Plan

This project is about conducting:

1- a statistical analysis by combining data from our git repos, the Open Build Service, Bugzilla, SMELT (the Maintenance database/app), the Support database, and our Sales databases to gain insights. Metabase is used as a BI tool to navigate the data

2- modelling and prediction of monthly/weekly bug report rate via SARIMAX models

https://www.machinelearningplus.com/time-series/arima-model-time-series-forecasting-python/

Results

Results are only available to SUSE employees and have been presented in the SUSE Manager 2019 Kickoff.

Slides, code and data are available here:

https://trello.com/c/txghSlKM/24-1430-l3-bug-review-and-statistic-development-team