GitLab Pipeline Authoring group, 25 Mar 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Value stream analytics - Pipeline Authoring

Description

https://gitlab.com/groups/gitlab-org/-/analytics/value_stream_analytics?created_after=2022-02-24&created_before=2022-03-25&value_stream_id=1284&label_name[]=group%3A%3Apipeline%20authoring

A

Okay, I want to caveat this heavily with. I have no idea what I'm doing, but I thought it was pretty cool. um I saw this this feature uh years ago and it was in his infancy, but like this is kind of useful. Now um there was a comment from from nadia in in one of the the retros. I can't remember what it said to be honest, but it was around how long do we spend doing something, and it got me thinking.

A

Maybe we should take a look at these value stream analytics and and see if they can help us right, so I've set up something really simple I'll give you a link to it, but basically um I've set up these five stages.

A

um Design plan breakdown, ready for development in devon in review and all it does is checks all our issues, for um you know how long were they in design and I'm defining that based on when the design label was added and the design label was removed. So it's the amount of time between these two things same for planning breakdown, ready for development in dev and in review, all sorts of other things you can do, but um I'm starting here I'm starting small because there's a lot of confusing data here.

A

To be honest, um there's all these things about tasks and stuff. I don't know what that is, so I'm going to skip over that uh lead time cycle time whatever.

A

So, if you go into um actually in dev, is a good one to look at for us um big spike here, where the average time for completion is 328 issue less days, um and we can see there. It is. uh I think we all expected this one to be. The outlier uh retrain failed uh I'll cancel job, so this has been in dev for almost a year which, which is rough, but it's good that this points this out right.

A

uh I assume this was only moved a minute ago.

A

I don't know oh okay, right yeah, so uh the other downside to to this is because in this case we got uh some uh levels wrong and and quickly changed them.

A

So as far as value stream analytics are concerned, uh this only spent one minute in in development, because it's the time between us, adding and removing the label, but whatever um it gives us kind of a bit of a view in what's potentially problematic and roughly how long we're spending in these things- and this seems about right to me- it's it's taking the median which is helpful for these outliers.

A

um We have a week in design 19 hours in planning breakdown, which is pretty good uh three weeks in ready for development yeah. I'm not sure how useful this one is because there's some issues that can sit and ready for development for a long time until they get prioritized um in dev about a week and in review about a week. These all sound about right, uh but yeah um I'll, send you the link. I have a play about, and let me know if there's there's anything useful in here.

A

If there's anything you want to add, if you do want to add something feel free to write it yourself, you've got permissions. Just go to edit here, just go to edit here there. It is it's a little buggy, if I'm being honest but hey it's getting up right. uh So you can add another stage.

A

Give it a name and then here's all the events you can have. So when the issue is closed, created added to a board associated with a milestone label was added levels removed.

A

Mr was closed, created the first commit first deployed to production, there's all sorts of stuff in here and then you can go through and add them in, and it just creates a new stage up here, which is pretty nice um again, not sure what this overview is. uh I would like insight into why this spike is here. Can I actually let me do that?

A

Oh this is nice and you can see. I am discovering this at the same time, everyone else is.

A

Oh, it doesn't.

A

Doesn't really change anything.

A

ah I just I don't know what this is. Whatever these ones are cool, though, uh to then you know, you've got a timeline of whatever. uh If we go back uh to a little start of the year, we can see roughly how well we're doing uh again, there's big spikes january 28th. Don't know why that is. I don't know if this will update to no, it will won't tell me um but yeah, it's still kind of in this infancy, but pretty cool.