[OtomotO]

(Needless) complexity is going on.

KISS and you sleep better.

That and the problem of forever chasing trends and never saying: "It's done" without reinventing everything every couple of years (trends again)

Sounds too easy? It is of course simplified, but the core still holds true.

GitHub just worked, but they had to migrate to React because "that's what everyone else uses"... Pure Enshittification.

[stabbles]

After yesterday's outage they admitted that their elasticsearch index for issues/prs lost data.

They seem to have changed the primary source of data in the issues and pull requests tabs (w/o filters applied) from the underlying database to the elasticsearch search index, which has the side effect that there's a noticeable delay between state change of an issue/pr and an update in the UI. But as seen today, these can get out of sync, and apparently they even had data loss in the index.

I would really like to know their reasoning for making that change. I can totally imagine that they wanted to "simplify" so the UI uses only a single data source instead of two.

As a user it's incredibly annoying to have a delay between issue/pr state changes and the search index picking it up.

[jrochkind1]

Yeah, I have been noticing weird things with Issues and PRs, including outdated state, for months now.

When the outage happened yesterday I sort of figured it was something I had been noticing building up or something.

[SoKamil]

What? React has nothing to do with current state of affairs. In fact, React on GitHub currently exists in mere islands, i.e. in Projects and recently in Pull Requests. Most of the frontend is still Web Components[1] paired with Turbo[2] for hot reloading. GitHub is still as slow even with JavaScript disabled, try it yourself. Backend just serves stuff really slow. In fact, there is an alternative GitHub frontend (no affiliation) that feels snappier and is written in React.[3]

With that said, Mitchell complains about outages. These started directly after Microsoft acquisition[4] and are attributed to migration from AWS to Azure.

[1] https://github.blog/engineering/architecture-optimization/ho...

[2] see html source for tags

[3] https://my.githero.app/

[4] https://damrnelson.github.io/github-historical-uptime/

[jrochkind1]

Pull Requests is the thing that was wacky in the UI yesterday, coincidence or not? I have no idea.

Yesterday we saw PR pages that displayed no error, just displayed wrong info. I would have preferred to get an error page than outdated or empty lists. I was guessing this was related to the React migration but I don't really know.

[queenkjuul]

I think the backend is just fucked. I have issues with Actions and the API all the time, not just the web UI

[hylaride]

Also, the browser back and forward buttons no longer work in pull requests when going between PR tabs (commits, checks, files changed, etc) as well as some other site interactions.

Like, what user-hostile intention was the reasoning behind that? I am literally imagining a product manager smoking a cigar and laughing at the RUM session replays of me losing my shit.