Discussion
Cascading failure. It cascaded.
> As unhealthy nodes dropped out of the global pool, traffic distribution across healthy nodes became imbalanced, amplifying the impact and causing intermittent availability even for regions that were partially healthy.
https://azure.status.microsoft/en-us/status/history/#incident-history-collapse-YKYN-BWZ
Oh don't worry, Azure is not having a *global* outage, just a "non-regional" one.
@rysiek another huge outage!?
@rysiek another huge outage!?
If you're a journalist writing about this Azure outage, don't forget to ask Microsoft how much of their code is "AI-generated".
Satya Nadella claims 30%:
https://www.cnbc.com/2025/04/29/satya-nadella-says-as-much-as-30percent-of-microsoft-code-is-written-by-ai.html
@rysiek “now written by software” could also mean things like the Azure SDKs which are auto generated (not by AI) from the API docs.
But, he probably won’t say that bit out loud because it’ll scare the investors
@rysiek Also, a regular reminder that lines of code is a shitty metric. That 30% number sounds way less impressive once you start seeing changes that have at least an order of magnitude as much code as needed just because LLMs make generating it so easy.
@marcink they themselves used that metric so I will gladly use that same metric against them in a context where "spicy autocomplete did most of the work" is Bad, Actually™.
@rysiek Oh yes, absolutely. It’s just that I’d love, also, for a journalist to respond to that 30% claim by asking about what exactly makes it a good thing.
Microsoft Azuren't
@rysiek ...Copilot 365...
...Xbox...
...Pro.
Microsoft Works .. doesn't
> We have confirmed that an inadvertent configuration change was the trigger event for this issue.
Folks, it's all okay! It was just an inadvertent configuration change [that happened to bring loads of services for millions of users down]. No biggie.
@rysiek
Totally on brand.
Bad configuration by Azure normally only exposes government emails to foreign actors.
@rysiek Next time it’ll be an advertent change, just to keep it fresh.
Look, this is the level of professionalism we should expect of Microsoft at this point.
They had a separate global outage of Office 365 (or whatever it's called this week) *this month* already:
https://www.bleepingcomputer.com/news/microsoft/microsoft-365-outage-blocks-access-to-teams-exchange-online/
And then there was the one from February this year:
https://www.upi.com/Top_News/US/2025/03/01/microsoft-outages/3061740867107/
Not to mention smaller ones, like the one in US-East that prevented admins from accessing the admin center:
https://www.bleepingcomputer.com/news/microsoft/microsoft-investigates-outage-affecting-microsoft-365-admin-center/
@rysiek
Office 351 or something
"or whatever it's called this week":
They should at least reduce the "365" by one, for every outage.
I wonder if firing 15.000+ employees just this year had any impact here. 
https://www.windowscentral.com/microsoft/microsoft-continues-layoffs-with-42-more-roles-cut-for-the-fifth-month-in-a-row
@rysiek What was the impact to Microsoft Sovereign Cloud and Microsoft Cloud for Sovereignty (two different offerings)? Because they're completely isolated from the rest of Azure, right?
@rysiek
Define inadvertent
Noun
Made by an AI after we fired the humans
@rysiek For some reason full PRs to dotnet and azure projects from MS employees seem to be mostly hidden. But here's a glimpse into the quality of the generated code and an example of how interacting with copilot impacts the development experience
@rysiek Looking forward to more outages.
A space for Bonfire maintainers and contributors to communicate
