@andr3@mnohr@phikai as usual, here's the Mural for us to add the issues we want to see in 14.0. I scheduled our 50-minute session for May 13.
Add your issues to the Mural before the call. Let's try to limit to 5 issues per person, so it's easier to vote on them and keep things focused. You can find instructions on how to add them in the “Outline” panel on the right side of the Mural UI.
Try not add Security or Availability issues. This is also noted in the product processes page, as those issues have forced prioritization with SLAs/SLOs.
If you can, mark issues that appeared in previous sessions by changing their sticky color to orange.
@andr3 I noticed that at the same time on May 10 there's am “Application performance session” in the Code Review group calendar, but it was not on your personal calendar. Should I reschedule the prioritization session?
@pedroms Thanks for asking. It was a good experience.
Going in I was a little unsure about what types of issues we wanted to talk about during the session. It seems like to focus of this session was to cover issues that may fall under the radar with the rest of our planning process. For example, we didn't talk much about the OKR FY22 Q2Large MR issues, because we know we are going to work on those. There are also issues from %13.12 that may slip, security issues, performance issues, etc. that will take priority over anything we discussed during this session.
What are the next steps for the 4 issues we identified as the "winners" of this process? We agreed they are the most important of the issues we discussed, but how do we now balance these 4 issues with all of the other work we want/need to do in %14.0. I think that is part of this process that I am still unclear on.
@mnohr apologies for the late reply. Thanks for your feedback!
Scope: Yes, you're spot on. There is work that we will have to do regardless (the kinds of issues you mentioned), and that's why we didn't talk about it during this session.
Prioritization: That's a great question! We raised these issues and discussed them because we feel they matter. There's no guarantee that they will be scheduled. At the end of the day, how they are prioritized over other work is @phikai's call, so I'll let him talk about it
What are the next steps for the 4 issues we identified as the "winners" of this process? We agreed they are the most important of the issues we discussed, but how do we now balance these 4 issues with all of the other work we want/need to do in %14.0. I think that is part of this process that I am still unclear on.
@mnohr it varies and was really my big question when I joined the group and we started this. In many cases we try to fit the items in and include them as part of planning. In other cases there's simply no room based on all of the other commitments we have.
I wonder if we should add a label to those so we can try and make sure we're actually getting through them.
I am going to do a first pass at data gathering for %14.0 here. On the Kanban board there are a number of other issues in the "Open" column that I am not sure what to do with yet. They will need more investigation.
14.0 Backend Capacity
w34 (Includes 10% Time)
13.12 Issues at Risk
These are Deliverable issues that have not been started yet and are at risk of not being completed in %13.12 (w15):
gitlab-vscode-extension#344 (closed) - This unblocks the ability to create new comments in VS Code and prevents a dependency chain on compatibility with GitLab needing a bug fix in GraphQL.
@phikai Yes, if we implement the workaround gitlab-vscode-extension#344 (closed), we'll get unblocked and we won't require GitLab version 14.0 or higher . Now that we realized that we wouldn't support older versions, it might be a better option to go with the workaround even though considered from the long-term it's a "throw away" work.
@viktomas I think for the sake of keeping this moving forward we need to go that route. The added bonus is we don't need a version with that fix. Do you think we can do the workaround and commenting in the same milestone?
Do you think we can do the workaround and commenting in the same milestone?
Yes! I'll start with the workaround and then proceed with the unblocked "new comments". The only requirement is that one is implemented after the other. I'm so excited! %14.0 is going to be a big one
That takes us to a weight of 8, do we think that's achievable?
I think that the new comments are "as good as done™" , gitlab-vscode-extension#168 (closed) might be a stretch considering my time off in %14.0 and tiny spillage in the form of gitlab-vscode-extension#341 (closed) but I'll do my best. I think that de-duplicating the overview tabs will improve the MR review UX.
@mnohr@andr3 I'm a bit concerned that we don't have any issues, even investigations up on the board with our Large MR OKR Label.
Can we please try to get issues created - even if it's the things we're going to investigate so that we can appropriately plan capacity against the other asks of the group.
w0 Further improve Total Blocking Time performance metric of Merge Request Changes page with large amount of changes gitlab#327635 (closed) (waiting for virtual scrolling to be enabled and measure the improvements)
w1* Unable to create multiline comments gitlab#322188 (closed) — i'm still waiting for a full report of the status for this, but it seems like it might be backend only.
With gitlab-vscode-extension!253 (merged), we removed the execa module from our production code. We now fully rely on VS Code for running our git commands and we shouldn't have any more security issues related to executing arbitrary code.
This is essentially brainstorming, so I'll be thinking about it the whole milestone
*Note 1: I've marked this issue's health as at risk because the frontend work is pending completion by the backend. The theory is there won't be front end work at all, but we might get into a situation where the back end completes late in the milestone, with very little time for the FE to ship if we do need to do something.
The main feature is merged in main I found one small edge case bug that I fixed and MR with this fix is going to be ready for review today. We can release this tomorrow . I created a new issue for a better UX when we fail to create a comment: gitlab-vscode-extension#398 (closed). @phikai@andr3 Should I pick it up in %14.0 ? sidenote: the comment doesn't get lost even with the initial feauture, you just have to get it from the logs
I had an idea for a feature that would allow you to create and apply patches stored in GitLab snippets gitlab-vscode-extension#393 (closed) Check out the video, I think it makes for a neat workflow and implementing the feature is quite straight-forward (frontend-weight2)
UPDATED - the sheer size of the merge requests table causes timeouts on staging, so back to the drawing board, and... 3 different DB engineers and no one really knows the path forward. This is slipping into %14.1 now, which pushes a number of other WIP -> Draft cleanup steps into %14.2.. I don't like it either.
I believe the backend change is mostly ready to go, but we have some edge cases where we make an additional full diffs request which I believe needs some frontend involvement. It's only an issue when we have files that have whitespace changes only, but I think we need to address that issue.
Thanks for the updates on Load merge request diffs incrementally from Gitaly - starting to look promising there so hopefully we can get some wins out of it!
Draft: Remove diffs gradual load feature flag. It's a FF removal which requires more attention due to many failing specs. I just picked it up as I noticed that it was default enabled several months ago. I took over an MR @iamphill started earlier and fixing failing specs in it. It was making diffs_batch code more complicated so I'm getting that FF removed while I'm at it.
I fell in love with TS again when introducing no-floating-promises eslint rule to the codebase. Requiring each promise to be handled is better than sliced bread.
GitLab Discussion API has sometimes troubles submitting slash commands in new comments (gitlab-vscode-extension#357 (closed), gitlab#326443 (closed)) I'll keep an eye on production logs to see if there's an influx of errors when creating comments. But so far it looks like roughly 1/day.