Google suffered hour-long wide outage in nearly every service, cause unknown
On Monday morning, a large percentage of Google service users worldwide were unable to use services while logged in, spanning YouTube, Gmail, Google Docs, Google Classroom, and more.
Google's Gmail abruptly went down some time around 6:30 AM Eastern Time, and presented users only with error messages. It was followed by similar outages on Google's major services, including Google Docs, and YouTube.
At around 7:30 AM Eastern Time, it updated its status pages to say the issue was being fixed for each service.
"Gmail service has already been restored for some users," said Google in the Gmail status page. "We expect a resolution for all users in the near future."
Similar messages were provided for other services through 8:20 A.M. Eastern time.
Gmail users using email apps can't get into their accounts
The flaws were related to authentication, making the problem expand past Google's services itself. Impacted users could not log into any third-party service through Google's authentication services -- including Google home automation.
Authentication has been restored to some users, but not all. It isn't clear what time full restoration will be completed.
Google's Gmail abruptly went down some time around 6:30 AM Eastern Time, and presented users only with error messages. It was followed by similar outages on Google's major services, including Google Docs, and YouTube.
At around 7:30 AM Eastern Time, it updated its status pages to say the issue was being fixed for each service.
"Gmail service has already been restored for some users," said Google in the Gmail status page. "We expect a resolution for all users in the near future."
Similar messages were provided for other services through 8:20 A.M. Eastern time.
Gmail users using email apps can't get into their accounts
The flaws were related to authentication, making the problem expand past Google's services itself. Impacted users could not log into any third-party service through Google's authentication services -- including Google home automation.
Sitting in dark like an idiot since #google is down and my Google home is in control of my lights. Now gotta download and use the actual app from wipro to launch my lights. pic.twitter.com/jhWNp7dkga
-- Auro (@weekendbiker)
Authentication has been restored to some users, but not all. It isn't clear what time full restoration will be completed.
Comments
I have no idea what that's supposed to be. Still, it's got a hammer so …
EDIT: Enterprise now clearing up too.
The customer also has to request the credit (it isn't automatically provided), it can only be used against future fees, and excludes outages due to "factors outside of Google’s reasonable control" (among other exclusions).
The short version: You will not get a credit, and even if you do, it is essentially worthless as it pales in comparison the the amount of money lost by a business as a result of these kinds of events.
Well, as long as there's a monkey on the case …
To pick on Spectrum, they are now advertising 99.9% network "reliability" for their business connectivity service. Okay, first of all they shouldn't be calling this "reliability" because what they are referring to is actually "availability." Ok, not everyone needs to understand the correct terminology and one could argue that availability is part of the reliability engineering domain. But what everyone does need to understand is what 99.9% availability really means.
Availability metrics follow a "rule of nines" rating system in terms of defining how much downtime is allowable to sustain the number of nines past the decimal place. Claiming that something meets 99.9% availability may sound good on a TV commercial while you're munching on Cheetos and guzzling Mountain Dew, but in terms of allowable downtime it equates to almost 44 minutes a month if the system is expected to be available 24x7x365. I'd bet that if those TV commercials said something like "we guarantee that you won't lose your service for more than 44 minutes a month" people would think they're getting pretty crappy "reliability" or "availability" depending on your understanding of this topic. From what I've heard about this latest Google service disruption and trying to recall recent disruptions, Google's service availability is probably still well above one-nines rating, which is generally considered to be acceptable for most business operations that don't promote themselves as "high availability." There are plenty of downtime calculators available online that use the rule of nines calculations.
So yeah, if you're a business that requires a specific availability performance to protect yourself from loss (perhaps in conjunction with an insurance policy) then you'd better make sure that you're standing up all of the right stuff to protect yourself, whether it's cloud based, local, or a combination. This may involve local backups, (various kinds of) redundancy, prepositioning resources, manual overrides, etc. There are plenty of industries and companies that require high availability and they go to extremes to provide the best defense against downtime and put survival techniques in place for getting over certain types of disruptions. It's very expensive and never perfect.
That's the price of paying for YouTube TV. (Pun intended)