Ocim containerization project(s) -- looking for epic owners

Ticket to log time: FAL-1874

We’re discussing how to split the SE-1834 Ocim containerization epic into smaller projects, so that we can spread the capacity and learning required for implementing this major change across all of the cells. It’s a fantastic DevOps opportunity – incorporating kubernetes, terraform, Tutor, Docker, and more. I’m stoked.

It’s a non-billable, non-cell project, so won’t hurt cell sustainability. It’s also community contributions, so can count towards core-committer time. But we still need to ensure that we plan and estimate the work carefully, to keep the scope tight, coordinate the work between cells, and parallelize what we can so we’re not blocking each other. Plus, capacity is still an issue, and so we also need to plan to bring newcomers into the work.

@adolfo is leading this project, and it dovetails nicely with chairing the Build-Test-Release Working group. He proposes that we split the project into these smaller epics:

  • Creating an image builder
  • Developing Tutorraform further, including making use of the output of the above image builder
  • Refactoring Ocim use the above and drop its equivalent features
  • Moving customers to the new infra (one epic per cell)

(I think the first 2 may be combined, but that’s a detail we can sort out after discovery/estimation.)

So what we need to know is:

  • Who’s interested?
  • Who wants to own an epic?
  • When (what sprint) can we book in the discovery tasks?

Hands up? :grinning_face_with_smiling_eyes:

9 Likes

Thanks for restarting this discussion, @jill!

The most pressing on the list, for now, are creating the image builder and improving Tutorraform (including renaming it, as requested by Regis). I think they’re safe to handle separately. Plus, it’s the only clear way of splitting the immediate work between epics that I could find.

The general discovery is this one (for which I had already asked for comments a few weeks ago). I recently rewrote it to be based on Braden’s work on Tutorraform, including his suggestion of The Three Major Pieces - hence why the first three items in my proposal.

Getting down to business, I’ve already started to tackle Tutorraform, and had planned to devote most of my time to it over the next few weeks (also as part of an initiative to focus my attention on a single project). However, Serenity is currently struggling with capacity, and unless we get one more newcomer beyond Sandeep for May, I won’t be able to. And given our recent bad luck with newcomers, that’s not likely to happen. Not to put too fine a point on it, I’m bummed.

As project lead, I’d rather have it move forward than wait on my cell’s capacity. But I’m not ready to give the work up entirely. :wink: I intend to not only propose that Ocim (and its accompanying deployment mechanism) be owned by a single person, I would like to be that person. And that’s only interesting to me as long as I can work on it myself.

So. There are really only two epics up for grabs that need to be worked on now. The image builder and Tutorraform. One I’d like Serenity to keep. Given @jill’s enthusiasm, it makes sense for the other to go to Falcon. Jill, any preference?

PS: I’m now filling in for Sid as Serenity’s recruitment manager, which goes counter to my objective of focusing on less things in the short term. But it sure puts me in a good position to solve the problem in the long run. :stuck_out_tongue:

I’m excited to see this! :rocket:

I would be happy to be a reviewer for one or two of these epics if there aren’t enough volunteers / cell capacity.

1 Like

This is an update to the thoughts I posted above. Something of a reversal.

As I stated previously, the main problem is that Serenity will not have the capacity to handle this on its own. It wouldn’t have the required sustainability, either, if it weren’t for the fact the epic is accelerated.

However, and this is what changed since my last post here, this sudden lack of capacity has had a large impact on my own sprints. I was already vying to reduce stress and improve efficiency by looking for something to focus on - and things have gone exactly in the opposite direction, unfortunately.

So, to keep it brief, I’m now looking for somebody to replace me as epic owner on SE-1834. (I’ll also be closing/relinquishing any planned or ongoing Ocim tasks, and will be dropping the devops specialist role, but that is for another thread.) I’ll continue as build-test-release chair, but only as a communication facilitator, so I’ll need whoever owns the epic to help me find assignees until Lilac is out. After Lilac, I’ll be opting out of release work entirely, and I suggest whoever takes this epic on should probably start attending the meetings.

Pinging @jill specifically: interested in taking it all on? :slight_smile:

1 Like

@adolfo, I’m interested, but I also don’t have time :sweat:

Unless I can hand over the LabXchange epics to someone else in @falcon ?

Like I said on the ticket I would be happy to take it @adolfo, even if I am making some experience with Ocim. I would like to learn more about it and it sounds like a nice opportunity.

My only concern is about the build-test-release role, I won’t be able to attend meetings and stuff like that, so I don’t really know if it is a blocker.

Thanks for replying, @jill, @toxinu.

It would be pretty important for whoever takes on SE-1834 to be in the loop on all matters BTR and Tutor. Ideally, that person will attend the bi-weekly meetings (including the Tutor maintainer’s one), but I can imagine having a lieutenant that can would also work. Which is to say, if you can find an epic reviewer willing to do that for/with you, it has a good chance of working.

I’d have suggested @jill be just that reviewer, but I figure the meeting times are just as bad for her as they are for you. Maybe @raul?

Will this project result in the replacement of Ansible usage in OCIM?

Yes. (And yes again. :slight_smile:)

1 Like

I’m interested, but (1) I don’t currently have time, and (2) I want to focus on fewer roles/epics and do a betters job at them once I finish some ongoing epics.

1 Like

I’m interested, and scheduled ticket to read all related discussions and try Tutorraform.

2 Likes

I’m also interested if 1) there is still not enough capacity and 2) there are well defined largish chunks of work instead of many small ones

I am highly interested. I don’t own any epic yet! And it looks like an amazing opportunity to learn about Kubernetes and scalable infrastructure in general. :slight_smile:

3 Likes

@shimulch, that’s great! Would you be able to attend the bi-weekly build-test-release meetings? Next one is on Monday at 15:00 UTC. It would also be best for the owner of SE-1834 to become a Tutor maintainer and attend those meetings as well.

If you’re comfortable with that and nobody else objects (I certainly don’t! :slight_smile:) , you can go ahead and assign yourself SE-1834.

2 Likes

Yes @shimulch and @demid, go for it! No objections from me either :smiley:

@jani You’ll need to wait until you’re confirmed on the core team to take lead on a project, but your help on the tasks in the meantime is very welcome.

Thank you all for putting up your hands!

1 Like

Yes! We need more people learning k8s, that’s amazing! :star2: :star_struck:

1 Like

@jill it was poor wording on my part. I definitely did not mean to volunteer for taking a lead but showed my interest for 3-5 story point tasks if any :)

2 Likes

@jani There will be big stories out of these projects, guaranteed. Thank you for voicing your interest and volunteering to help! :smiley:

I’m also interested and have a few years of experience running production stuff on k8s, in case more help is needed :slight_smile:

1 Like

@adolfo, @jill, I’ve created a ticket to learn about Tutorraform as well.

Regarding the split, would this be a better idea to close the current epic - SE-1834 and creating new ones?

@demid Do you have a preference on which epic you want to take on? I mean Tutorraform or the OCIM end? :slight_smile: