nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

how is total age calculated in average age report

Sanne van Tongelen

Every week I have a look at the average age report, use the same filter, timeframe and lookback window. I noticed the data of previous weeks changes, this seems incorrect, as those weeks have past, so the data of these weeks shouldn't change.

For instance this week the average age of last week is 118, but when I looked last Monday it was 105.

Even if items that were on the backlog last week the average of that week shouldn't change today, as the average of the issues last week shouldn't go further than last week.

Am I missing something here?

Accepted answers

J van Leeuwen

Hi everyone,

I was able to find some information on how it's calculated:

The Average Age Gadget displays a bar chart showing the average number of days that issues have been unresolved. For every issue, we:

extract the creation date and resolution date
iterate over a list of contiguous time buckets (configurable by an hour, day, month or year)

if:

the issue creation date is before the start of the period and the resolution date is after the start of the period

We then compute the amount of time the issue has been opened up to the end of this period using min(resolutionDate, periodEndDate) – issueCreationDate and increment a count of open issues for that period.

The result is a map of periods to total time that the issues have been opened. We compute the average by dividing the total time by the total number of issues recorded per period.

This gadget does rely on the org.apache.lucene.search.Collector implementation. It's possible if there were some index inconsistencies that were solved by a later reindex, that this could explain the difference in numbers.

If you want to dig deeper, the code can be found in https://stash.atlassian.com/projects/JIRASERVER/repos/jira/browse/jira-components/jira-plugins/jira-gadgets-plugin/src/main/java/com/atlassian/jira/gadgets/system/AverageAgeChartResource.java#41,48,60,76,121.

I hope this answer is helpful.

- Jennifer

All comments

Jack Brickey

@Sanne van Tongelen, I have always found the chart to be difficult to consume but TBH i have not spent a lot of time recently trying to decode the math. I would be interested in your observations week-over-week. Specifically, do you see that week-over-week the the bar changes for several weeks back (not just last week). I don't know what you mean by "total age". My understanding is this chart is trying to illustrate if you are keeping up w/ you backlog or not. If the trend is climbing then you are falling behind if declining you are making positive progress and if level then you are treading water. I haven't monitored the number much.

Sanne van Tongelen

Thanks for your response.

I understand that it is meant to see if you keep up with your backlog.
We weekly keep track of the average age to see how we are doing over time and motivating the Product owner to keep the average age at a certain level. However we found the data to be inconsistent, at least to what we expect. Maybe we are expecting the wrong thing.

I added the numbers in our external KPI dashboard and in week 35 and 36 I had the numbers: 117 and 118.
However if I have a look at the chart today it has changed to 122 and 121:

In my opinion that wouldn't be possible, as that is changing the past. Today's stories shouldn't change the average of past weeks as the average back then was the average on that point in time.

Hopefully this creates clearance on my issue.

J van Leeuwen

Hi everyone,

I was able to find some information on how it's calculated:

The Average Age Gadget displays a bar chart showing the average number of days that issues have been unresolved. For every issue, we:

extract the creation date and resolution date
iterate over a list of contiguous time buckets (configurable by an hour, day, month or year)

if:

the issue creation date is before the start of the period and the resolution date is after the start of the period

The result is a map of periods to total time that the issues have been opened. We compute the average by dividing the total time by the total number of issues recorded per period.

I hope this answer is helpful.

- Jennifer

Sanne van Tongelen

Thanks for your reply, this clarifies the calculation.

Jack Brickey

@J van Leeuwen , thanks for this info. I'm wondering if this is detailed in Atlassian documentation and if not I wonder if maybe turning this into an article in the Community would be a good idea?

J van Leeuwen

Hi @Jack Brickey - You're very welcome. Absolutely, it would! That's next on my list. Glad it was helpful, @Sanne van Tongelen !

Jennifer

Senya_qua4dy

Thank you. How do I get to the Data Table?

Welcome!

It looks like you're new here. Sign in or register to get started.

Quick Links

Find more posts tagged with

jira

cloud

jira-cloud