Short version
I am consistently seeing a different number of (work)days in transition status for individual issues than what I calculate from the History in Jira. How does eazyBI calculate these values? Anything I may be missing?
Longer explanation
I’m trying to create an average cycle time report by component that EXCLUDES the backlog and done statuses, and an average lead time report by component that INCLUDES backlog but NOT done. I’ve created two new Transition Status calculated members to do this, but I am still very new to MDX.
On the surface the data look approximately correct, but when I try to spot check using Drill Through Issue the values there don’t match up to the History values for work items in Jira. They’re usually off by about 2 - 10 days if I look at working days, and more than that if I look at calendar days.
I noticed your question has not received any answers yet. I hope you have already found the solution; if not, there are some ideas.
Days in transition status are calculated based on issue change history, counting together all the periods when the issue was in this status. The average is calculated by dividing the total duration in the status by the transition from the status count. For workday calculations, time spent in the status during weekends and defined holidays is excluded.
It means, if the issue was in In Progress for 5 days, then moved to done, then reopened and got into In Progress for 3 additional days, the average will be 4 days ((5+3)/2). It gets more complicated when you calculate the average based on aggregated statuses (total duration in all statuses divided by transition count).
The ideal solution is using issue cycles, where you define which statuses must be included and then the cycle time is calculated automatically in days or workdays, and average as well. Read here: Issue cycles
However, currently, each status can be added only to one cycle per account, which is insufficient in your case. We have a feature request in our backlog to allow including statuses in more cycles. I added your vote for it.
A rough solution still is using aggregated transition statuses together with Average days in transition status or Average workdays in transition status, like in this thread: Average Lead and Cycle Time Report - #2 by lauma.cirule
Thank you for your reply! I eventually got to an answer very similar to the one shown in the thread you linked. It is good to know about Issue Cycles though as this may be a simpler way to get at some of the data we’re looking for.
