Again? This is the third time this month. The same thing happened just recently but for the whole of GitHub just 3 days ago: [0] Most likely related to the Microsoft 365 Azure outage. [1] This really doesn't look good for GitHub.

Self-hosting a backup sounds like a more sensible option than 'going all in' on someone else's server or 'centralising everything' on GitHub, as I have once again said before.

[0] https://news.ycombinator.com/item?id=26439436

[1] https://news.ycombinator.com/item?id=26469142

FWIW I'm self-hosting a Gitea instance on a server <200 miles away from me. I thought it'd be a stupid idea, but setting it up properly was a lot of fun, and it's just stupid, crazy fast compared to anything managed.

My clients mostly ask for Github but for my own products I use self hosted Gitea with backup and parallel server just in case

It's almost like everything becoming more centralized leads to more outages, complacency and status quo that slows down innovation and reduces redundancy.

Kinda, but based on ye olden times experience, we were all much worse at running our own hardware.

I will agree that despite the relatively high reliability, I think we'll likely have a black swan event at some point in the future that we will regret not having tested off site disaster plans.

How does this has anything to do with centralization? These recent GitHub outage largely linked with Azure outages, it is either they are migrating to Azure or the migration caused Azure outages

Large tech companies tend to buy other companies when they grow and with that, centralize the ecosystem. Microsoft buying GitHub and moving all the infrastructure to Microsoft is a classic example on centralization. And the frequent downtimes are another classic example on the drawbacks of centralization.

[citation needed]

I had an issue like this with Power Automate Desktop yesterday and today. It was trying to retrieve my automation script from their servers I'm sure and there was no response.

Hacker news, azure, github.

The internet gods are angry today

the worst part is that even the self hosted runners are impacted so it is likely that their scheduling system is down

I'm having issues on Heroku right now with git deploys not getting the right and updated code set. Took me far too long to notice the warning bar at the top of the page.