Where are we now, Microsoft 362.5? Europe reports outages Some Microsoft 365 users in Western Europe earlier today were reporting service problems today, though those in the UK seem to have avoided many of the issues. A scan of comments from users on Reddit show a range of complaints from error messages while trying to open Outlook to outages to LinkedIn being down. Users in Sweden, …
these companies.
They promise 99.99999% uptime but deliver a lot less. (The jury is still out on that one as to what it actually is)
Putting all these services 'somewhere' and on 'someone else's kit' is a beancounters wet dream but for the rest of us they are an absolute nightmare.
By all means give it a go but you may want to get your numbers right first. They promise 99.9% uptime on a monthly basis, so they're allowed 43 minutes of downtime per month **per service**. Then the terms (which you've accepted by signing up of course) tell you that it's only a financially backed guarantee so you get a percentage of your bill back.
I think you're probably better off trying to file a lawsuit against your beancounters if it's that detrimental to your business.
I do wonder how it really compares to self hosted.
10% of the world down for a few hours at the same time is more newsworthy than 50% of companies having 1 day outages at different times for different reasons spread across the year.
When an outage hits a smaller company they rarely have the ability to get back online in a couple of hours.
Yes, exactly. In my many years of experience with on-prem systems, I never saw a single one that could boast of 100% reliability over any medium or long term timeframe. But when ABC Co's internal network goes down, they don't post the details on a public status page like cloud providers do, and their employees don't take to Twiitter to bash the IT department. I'm certainly not suggesting the cloud is always better than on-prem -- it is not -- but I bet that the major cloud providers' overall uptime is way better than almost all on-prem facilities.
it's a single service that is down. Not everything
Previous orkplace the aircon for the (new) server room was wired to the lighting circuit for the floor. Which got shut down once a week so that work could be done. The aircon wasn't configured to come back up automatically..
When we got in in the morning, the air temperature in the server room was 65C. The servers (Sun [1] mostly) were struggling, but still up. One Sparc disk array 1 had fallen over but, once things cooled down, we powered it back up and had only lost 2 drives out of the 20 or so in the array.
Next maintenance window, the aircon was put on a dedicated circuit with a big "Do not switch off" label taped over the switch. *And* the aircon was reconfigured to default to powering up automatically..
[1] Yes, it was that long ago. Pre-y2k from memory.. or maybe just post-y2k
When I worked for a subsidiary of a huge US globocorp, our entire network was once taken out for three whole days because someone hit reply all to an all staff email, which prompted someone to hit reply all to ask people not to hit reply all, etc...
The ensuing storm filled up all the mail servers and took the global network down while the network team deleted the emails.
I wonder how the likes of Gmail or Office 3xx would cope with a reply all storm?
I do wonder how it really compares to self hosted
Has a better uptime than us currently..
(LSS - 4 chillers in the server room. Due to lack of maintenance [1] two of them don't work. A day or so ago, the other two turned themselves off. One managed to come back online when we power-cycled it, the other one is deader than Boris' political career..)
[1] We are not allowed to do it - it's the responsibility of Building Maintenance. And, apparently, fixing the server room aircon (or even making sure it's serviced regularly) is *expensive* so it hasn't been done.. Well, it's a damn sight more expensive doing an emergency callout for an aircon engineer then rapidly buying two mobile units when you discover that they have a 2-day lead time as well as having the entire dev team sit around doing nothing because their servers can't be bought up because the 1 working aircon unit wouldn't be able to cope and will go into thermal shutdown again... I think we'll be taking over the aircon maintenance from now on.
"I do wonder how it really compares to self hosted."
When self-hosted you know which engineer to call and not be in a queue with hundreds of other O356 victims waiting for some page turner to "help' you before escalating. (With optional survey at the end, just press #2)
When MS came out with the Azure Cloud while limiting options for on-prem, no one objected and they all deserve what they get.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
