We are pleased to announce that we are evaluating a shift in our approach to performance measurement for Jira Software Data Center and Jira Service Management DC products. The main concept is to transition from a regression-based model, in which we compare performance against previous versions to detect and address regressions, to an NFR compliance-based model. In this new approach, we aim to identify performance NFR violations against specific thresholds and concentrate on resolving them across different product areas. Moreover, we will be adjusting the toolchain and utilizing various data shapes that more accurately represent our customer instance shapes.
It's important to note that we are still at the initial stages of this process and would highly value your input regarding what you would like to see in our performance reports. You can also provide feedback by reaching out to me at dmohanasamy@atlassian.com.
Thank you for your cooperation.
Sincerely,
JSW DC and JSM DC team
Hey Martin,
In the old model, we were comparing the performance levels with earlier versions, and working on fixing the ones whenever we identified a drop, even if both the levels were in the already acceptable range that customers wouldn't feel any perceived performance difference.
In the NFR compliance-based model, we are going to be comparing performance levels against pre-identified threshold levels. Example: If we identify that TTI for an issue view is 0.8s, a future performance drop from 0.4s to 0.3s after we have introduced a capability wouldn't warrant a lot of investment from our side to fix them before we can ship this new feature.
It's just one aspect to it. The ease of testing, better datasets, new toolchain and addition of other measures are at play in this initiative.
I hope this clarifies.