One thing I hate about GH's status page: sure "auth" gets to live under API, so API gets the downtime report today. But all of my remote git ops are failing, because i was dumb enough to use the `gh` cli to auth. Their status pages are not well defined, but they treat them as such, leading to inflated uptime numbers.
They've been moving it to Azure, presumably partly because they don't want to pay AWS for a thing they could (apparently) do, but more for the appearance of the thing.
I think that's financially sensible in the long run, assuming they believe they can run the thing basically as well on Azure, and aren't expecting major issues during the migration. These have turned out not to be good assumptions.
There is ALWAYS a need to scale up, upgrade versions, migrate from one database or scheme to another, cut services into more microservices, melt them back into a monolith, integrate more stuff, deprecate other stuff, fix bugs, increase performance, tune caches…
Some things happen for reasons that make sense, others are done for the sake of it. The PO‘s of front end features need to evolve them. What do you do with a PO who says this is done? They don‘t have visibility, don’t get promoted and they don‘t get a bonus. We have created a dashboard that shows that A/B tested users like the new design more by 7%! The next redesign is already planned!
Line must go up. They're already the default choice for hosting software repos online, so it's time to look at the next thing, whether that be CI/CD or AI or whatever.
If line doesn't go up, some Ivy League business school grad who has holdings in MS might make slightly less money and have to cut back on the nose candy budget while summering in the south of France this year.
They need to split paying customers off onto dedicated infrastructure. I know that you get that with some enterprise offerings but I don’t believe the normal paying customers get that treatment.
It’s clear they have massive scaling challenges but to not tier or isolate your paying customers from the free plan users seems a bit odd. Just my 2cents though.
To a certain extend I can fully accept that GH has scaling issues because of the insane traffic increase from AI generated code. But they are owned by one of the largest technology companies in the world, who also happens to run a hyperscaler.. so..
If I had a nickel for every time GitHub had issues, I'd be able to afford a nice stick of DDR5.
What are you going to do with your 4GB of DDR5?
Ah so that's why the GitHub Pull Requests extension is repeatedly harassing me about needing to log in again...
Or the GitHub mobile app where I was using a remote copilot session.
I just set up Forgejo on one of my machines. It works very well, have not had any problems yet.
Race for the mythical eight eights of uptime continues!
One thing I hate about GH's status page: sure "auth" gets to live under API, so API gets the downtime report today. But all of my remote git ops are failing, because i was dumb enough to use the `gh` cli to auth. Their status pages are not well defined, but they treat them as such, leading to inflated uptime numbers.
I'd much prefer to get a 500 than a 401 if there is something broken in the server. This wasted a solid hour of my day.
Surely they just need to consume more tokens to fix this!
These are not working for me right now: Github Pull Requests VS code extension, Refined Github Chrome extension
neither for me
This is what I get for trying out Fable 5.
Still having issues codeql actions error out with auth issues. :(
cool looks like this nuked all of our authenticated connections in Heroku, too.
I thought I was hacked this morning... Nope just Github being github...
GitHub iOS logged me out. Tried logging in twice before realizing the problem is on their side...
CodeQL errors with auth issues brought me here.
Thankyou, was just starting to dig into what was going on with my build
Why are they even changing GitHub?
Like what is the deal here? Why is the company ruining this program? Surely there can't be a financially sensible reason?
They've been moving it to Azure, presumably partly because they don't want to pay AWS for a thing they could (apparently) do, but more for the appearance of the thing.
I think that's financially sensible in the long run, assuming they believe they can run the thing basically as well on Azure, and aren't expecting major issues during the migration. These have turned out not to be good assumptions.
Because of vibe coding everyone is a developer now so the number of accounts, PRs, repos, etc just 10x'd.
There is ALWAYS a need to scale up, upgrade versions, migrate from one database or scheme to another, cut services into more microservices, melt them back into a monolith, integrate more stuff, deprecate other stuff, fix bugs, increase performance, tune caches…
Some things happen for reasons that make sense, others are done for the sake of it. The PO‘s of front end features need to evolve them. What do you do with a PO who says this is done? They don‘t have visibility, don’t get promoted and they don‘t get a bonus. We have created a dashboard that shows that A/B tested users like the new design more by 7%! The next redesign is already planned!
Line must go up. They're already the default choice for hosting software repos online, so it's time to look at the next thing, whether that be CI/CD or AI or whatever.
If line doesn't go up, some Ivy League business school grad who has holdings in MS might make slightly less money and have to cut back on the nose candy budget while summering in the south of France this year.
And that's just inexcusable.
Again? Last time an incident happened was 48 hours ago [0]. Now the API is having issues.
Something every week is falling apart at GitHub and it is slowly self-destructing.
Each time there is an incident at GitHub, it is another advertisement to self-host your own. Otherwise just expect GitHub to break next week or two.
Remember, there is no CEO of GitHub; only their AI agents running it into the ground.
[0] https://news.ycombinator.com/item?id=48442572
OpenAI, Anthropic and Google serve models, but approximately all of their combined output ends up on github.
They need to split paying customers off onto dedicated infrastructure. I know that you get that with some enterprise offerings but I don’t believe the normal paying customers get that treatment.
It’s clear they have massive scaling challenges but to not tier or isolate your paying customers from the free plan users seems a bit odd. Just my 2cents though.
To a certain extend I can fully accept that GH has scaling issues because of the insane traffic increase from AI generated code. But they are owned by one of the largest technology companies in the world, who also happens to run a hyperscaler.. so..
Nothing but respect for the GitHub team - they're at the center of it all. Can't imagine how their traffic looks these days.
They should raise their prices!