So is one nine 90%, or 9%, or 0.9%, or 0.09%, or 0.009%?
GitHub appears to be struggling with measly three nines availability
Scarcely a day goes by without an outage at a cloud service. Forget five nines – the way things are going, one nine is looking like an ambitious goal. GitHub has had a rough month so far. On February 9, Actions, pull requests, notifications, and Copilot all experienced issues. The Microsoft tentacle admitted it was having …
-
-
Tuesday 10th February 2026 14:25 GMT JohnSheeran
Well, assuming you aren't kidding.
99.9% = Three 9s of availability which means ~8 hours, 46 minutes of down time annually (this is terrible)
99.99% = Four 9's of availability which means ~53 minutes of down time annually
99.999% = Five 9's of availability which means ~5.3 minutes of down time annually (this is the gold standard)
No one is going to give you one or two 9's of availability as an SLA.
-
-
Wednesday 11th February 2026 11:01 GMT af108
No one is going to give you one or two 9's of availability as an SLA.
Well they might not give it to you in an SLA but that doesn't mean it won't actually happen.
Some organisations accept that breaking an SLA and paying for that is often "cheaper" than fixing the real source of the problem. It's an uncomfortable truth.
-
Wednesday 11th February 2026 17:17 GMT Elongated Muskrat
Microsoft might not be offering two nines as their SLA, but that's what they're delivering according to the "missing" status page. Not one single component of their service is currently meeting three nines.
I couldn't care less if their Copilot is at 0 nines, but their outage on Monday meant that it took several attempts to perform basic git operations. The website showed the chintzy unicorn when trying to view branches in the repo I was working on, and simple git operations such as pulling the latest version of a branch down to my development environment in Visual Studio (from an Enterprise subscription to github) were getting various HTTP errors (500 mostly). How you manage to break something as basic as a git server in this way is beyond me, and for a company the size of Microsoft to drop the ball this badly should be raising several pointed questions about why their service is so bad.
-
-
Tuesday 10th February 2026 16:01 GMT Anonymous Coward
Nine HFTs
Years ago I worked on monitoring distributed transactions. Some customers were high frequency traders, so scale and latency and correctness and availability were all important. They were all colocated with their markets, or right next to them.
A product manager arrived to teach us lowly developers about the Real World. The first thing he decided was that we needed availability of 5 9s, yay! But that we would measure it daily. The systems we monitored were only active during trading hours, making any measurement across 24 hours meaningless. He would not budge, not even when we told him that - at these time scales - even 9 5s would be too much. 9 5s is approximately 5/9, which across 24h is 13h20m - longer than trading plus pretrading.
-
Tuesday 10th February 2026 20:45 GMT sarusa
Microslop gonna Slop
Github has gone full Microslop, so with their new infatuation with vibe coding and vibe administration it's no surprise they're suddenly as unreliable as their parent company.
Don't expect this to get any better, ever, till the bubble bursts - if even then. I fled to GitLab.
-
Tuesday 10th February 2026 23:43 GMT Anonymous Coward
Économique avec la vérité
Actions, pull requests, notifications, and Copilot all experienced issues
When Microsoft say that somehow it also meant 9 times out of 10 the front page gave a 500 error and if it didn't then when you clicked sign in then it gave a different error involving a pink unicorn.
Absolutely absurd for a business to outsource their source code repo to Microsoft. I wonder how many times it'll have to break before people realise and it's all brought in house again.
-
-
Wednesday 11th February 2026 06:23 GMT Claude Yeller
Re: Mirroring
Making backups of important data, code, and services is ALWAYS a good idea.
Codeberg has already been mentioned. It is not in the USA and Free. Importing code from GitHub is easy.
But Sourceforge and GitLab also work well[1].
[1] YMMV! If you use GitHub actions etc, you will be toast, though.
-
-
Wednesday 11th February 2026 00:05 GMT Anonymous Coward
Did GitHub go the same way as Hotmail ?
I remember meeting someone from Hotmail back in the day.
They were taken over and told to migrate.
All those reliable BSD servers had to run on windows !
And to add insult to injury they had their own email migrated to exchange.
(I'm old. Memory may be wrong about details. Do not sue me.)
-
-
-
-
Thursday 12th February 2026 15:33 GMT O'Reg Inalsin
Re: It's simple...
After acquiring GitHub in 2018, Microsoft mostly let the developer platform run autonomously. But in recent months, that’s changed. With GitHub CEO Thomas Dohmke leaving the company this August, and GitHub being folded more deeply into Microsoft’s organizational structure, GitHub lost that independence. Now, according to internal GitHub documents The New Stack has seen, the next step of this deeper integration into the Microsoft structure is moving all of GitHub’s infrastructure to Azure, even at the cost of delaying work on new features.
In a message to GitHub’s staff, CTO Vladimir Fedorov notes that GitHub is constrained on capacity in its Virginia data center. “It’s existential for us to keep up with the demands of AI and Copilot, which are changing how people use GitHub,” he writes.
...
“We have to do this,” Fedorov writes. “It’s existential for GitHub to have the ability to scale to meet the demands of AI and Copilot, and Azure is our path forward. We have been incrementally using more Azure capacity in places like Actions, search, edge sites and Proxima, but the time has come to go all-in on this move and finish it.”
GitHub has recently seen more outages, in part because its central data center in Virginia is indeed resource-constrained and running into scaling issues. AI agents are part of the problem here. But it’s our understanding that some GitHub employees are concerned about this migration because GitHub’s MySQL clusters, which form the backbone of the service and run on bare metal servers, won’t easily make the move to Azure and lead to even more outages going forward.
New Stack, "GitHub Will Prioritize Migrating to Azure Over Feature Development ", Oct 8, 2025
-
-
-
-
Thursday 12th February 2026 16:31 GMT Elongated Muskrat
If you follow the link to the "missing status page" you can see for yourself, that the time period was October, with a measly 89.9% uptime. A total of 14h 37m of major outages, plus a whole load more of degraded service according to those figures. Although, to be fair, the way those figures are aggregated, those include outages in Copilot, which nobody in their right mind would care about. Automated AI reviews of PRs can be marginally useful, but they're hardly essential.
-
-
Wednesday 11th February 2026 15:34 GMT Stu J
Five nines is usually overkill
It may be a "gold standard" but for the vast majority of cases, the extra engineering complexity and cost to actually realistically achieve five nines (which is 5.26 minutes downtime per year) is a false economy compared to four nines - which is 53 minutes downtime per year.
I would go so far to say that - on paper - for 3rd party hosted SaaS products that aren't classed as "mission critical" (which arguably 365 and GitHub aren't really) if you get three nines you're probably doing OK - that's a little over 8 hours downtime in a year.
The caveat to that is that nines alone aren't necessarily a useful measure of lost value and disruption. Take three nines; if you randomly lose an hour's service here and there scattered throughout a year, the disruption to you extends beyond the outage window as you're playing catch-up repeatedly. You're probably more disrupted overall than by a single 8 hour outage, when you could just decamp to the pub instead.
-
Monday 23rd March 2026 13:51 GMT ChrisMarshallNY
I Have Always Considered "Five Nines" to be 99.999% B.S.
I've been seeing this stat, since the last century. Just about every hosting/SaaS/communication service that I've used, in the last thirty years or so, has claimed "99.999% Uptime."
Very few have actually delivered anything close to it. I have always assumed that they "game" the numbers, by excluding certain types of downtime.
