Slashdot Mirror
Microsoft's Multi-Factor Authentication Service Goes Down For Second Week in a Row (zdnet.com)
Just over a week after a global problem with its multi-factor authentication (MFA) service plagued a number of users, another Microsoft MFA outage is impacting a number of customers. Many, but not all, of the customers reporting problems today seem to be U.S.-based. From a report: Starting around 9:15 a.m. ET, a number of Office 365 customers began reporting on Twitter that they were unable to sign into that service because of an MFA issue. Office 365 is one of a number of Microsoft services that uses Azure Active Directory MFA to authenticate. Around 10:15 a.m. ET, Microsoft's Azure status dashboard was updated to reflect the possibility of a cross-region potential outage impacting MFA. "Impacted customers may experience failures when attempting to authenticate into Azure resources where MFA is required by policy. Engineers are investigating the issue and the next update will be provided in 60 minutes or as events warrant," the dashboard status said.
Yes: this is what happens when you don't own your software, you just "license" the use of it.
http://www.geoffreylandis.com
Locally installed applications are not exposed to this mode of failure. This story is about as interesting as people who complain about breakfast hours at restaurants. Cook your own breakfast any time of day.
Cue Airplane "They bought their tickets. They knew what they were getting in to. I say, let 'em crash."
Oh... THAT is what the "MF" in "MFA" stands for! I thought it was something else!
"Engineers are currently in the process of cycling backend services responsible for processing MFA requests."
So, they're turning it off and back on again.
"Freedom in the USA is not the ability to do what you want. It is the ability to stop others from doing what THEY want"
Die Like Lemmings
You have critical applications they have no business being in the cloud. Especially not someone else's cloud.
If no one can log in with MFA, no one can be hacked, can they?
https://www.foxnews.com/travel/life-sized-replica-of-noahs-ark-will-sail-to-israel-says-dutch-carpenter-who-built-it
ps: Azure failures are not news! Probably it was caused by a Windows update gone wrong.
Microsoft software that has only failed once a week? That HAS to be a record!
Choose how you want to run IT.
If you think you can run to the cloud and get better service you are mistaken. Like playing musical chairs you only move the problems and goal posts around.
There is no end to Management willing to pay through the nose for the promise of "Cloud" and following the advice of the providers along the way with little question, but when you have to build it on-prem you have to justify every blithering dollar you ask to spend and then have to face them trying to screw up your project plans with scope creep and "know-it-all" management interference and second guessing junior idiots.
In short, your shit is going offline... you want that reduced? Find quality IT pros and fucking pay them what they are worth and stop promoting high quality pro's to justify giving them a higher salary. If you need too... pay a helpdesk worker that gets their fucking shit done twice what you pay the others. It's that simple and stay the fuck out of their way... they are the professionals... not the fucking management. Managements ONLY job should be to make sure that money is wisely spent by make sure the teams are aware of talent and licenses product are not unnecessarily duplicated and that the nerds or silo managers are not busy fighting like children over stupid shit between themselves or other teams. Those are two huge problems but get very little attention in many businesses.
Cue the Microsoft blues (azures)
Sounds like someone at Microsoft tried to "improve" Azure security and it passed through MS's Quality Control team. Oh, wait! End users are the QC team.
Which is not a bad thing, in a world of constant change. In addition, if you get Microsoft products, you know what to expect.
About every 12 to 18 months, the owner of the company I work for will come to me about moving 'everything' to the cloud. I always say the same thing, "Maybe we could move {a few non-essential things} and see how that goes, but I wouldn't trust moving {anything we rely on}".
This article and many others like it are the reason I will keep saying this.
Welcome back to the glorious past of mainframe computing, where everything you do is in the hands of some big company.
I see all of those tedious little Personal Computing rebels, who wanted to run programs and keep data under their own control, have finally been exterminated.
"By the grace of Microsoft, do we compute".
Consumertards.
My workplace uses MSA for our VPN (which you have to be on for admin access to the servers). I'm starting to miss the RSA SecurID fobs we used to have.
Is still humming along with all my data very nicely. Have fun with that Office 365 outage.
cross-region potential outage impacting MFA
The whole point of being in the cloud is so if one region goes down you can switch over/fallback to the other region's servers to maintain uptime!!!
Heck, they cannot even keep the single level signon needed to connect to the STORE and Feedback application working. And downloads through the MS infrastructure have been stalled for a while. Nice to lock everyone into this mechanism -- provided you keep it working. But no... and the advise as to how to fix things always ends up with wipe the machine and do a clean install and reload all your applications. Maybe it will work this time... for a while.
I suspect their cleverness at setting up complicated mechanisms within the code greatly exceeds their ability to understand how it actually works and even more their ability to patch things. And since management is focused on deploying more and more features without a concern that key existing features don't work -- this is heading for a crash. Guess they should not have laid off their QA folks and outsourced development...
To all you ops guys who think no one can run infra as well as you:
Please stop the I told you so crap. For every one of you power-wizards, there are 100 fallible ops guys sitting in other chairs. Trust me, I've worked with a bunch of them over the last 40 years. Cloud platforms have outages a lot less than all the custom shops I've worked in, and I've worked in both big and small. Sure, Microsoft's outages are bigger and affect more people, but any particular company has only so much stuff that gets impacted.
Give it a rest - make the world a slightly better place.
- The Kessel run is for nerf herders. I can circumnavigate the entire Central Finite Curve in a lot less than 12 parse
... from accessing a host of internal applications at the company I was contracting with last Spring. And the internally-written authentication application was being slowly phased out and more internal applications were being migrated over to use the Microsoft application. By now, I expect that most, if not all, of those employee services were nicely locked down by Microsoft. One of these days, managers (and bean counters) will learn what is meant by "single point of failure".
CUR ALLOC 20195.....5804M
Comment removed based on user account deletion
Comment removed based on user account deletion
this is what happens when you force your software engineers into on call duties as support engineers. you have a bunch of people doing a job they werent trained for badly and neglecting their real jobs resulting in the windows os quality you have recently observed. 2 fails and bad outomes for customers of both
Safety ? you don't need THAT feature !
aaaaaaa
The whole point of the cloud is not to have good service.
The whole point of the cloud is to hand over your data to a third party, and to NSA, not to let your users acess it.
aaaaaaa
Ah, but you get to throw up your hands and blame/complain about someone else so the CYA factor is through the roof