back to article Worldwide Google services – from GCP to G Suite – hit with the outage stick

Google cloud services across the world are having a nap, with multiple GCP products including App Engine, Cloud Storage, Cloud Logging and BigQuery down since 06:45 UTC (07:45 UK time). The main issue appears to be with cloud infrastructure components, and Google Kubernetes Engine is also having problems. GKE clusters using …

  1. Pascal Monett Silver badge

    Oh my, Google Cloud and Slack at the same time

    It seems that all this cloudy stuff is really, really hard to keep working.

    We used to talk about five nines availability - seems like that's last millennium's history.

    1. Anonymous Coward Silver badge
      Flame

      Re: Oh my, Google Cloud and Slack at the same time

      It's Storm Ellen making it rain!

    2. Arthur the cat Silver badge
      Trollface

      Re: Oh my, Google Cloud and Slack at the same time

      We used to talk about five nines availability - seems like that's last millennium's history.

      They decided to improve it to nine fives.

    3. James12345

      Re: Oh my, Google Cloud and Slack at the same time

      GSuite manages upwards of four nines usually, but falls just short of five. Generally you don't notice the small wobbles it has.

      I very much doubt most businesses ran an email system with greater uptime, back in the days when running your own email system was seen to be a good idea.

  2. Luke Worm
    Happy

    Smiley

    Aah, the wonders of having your systems running on someone else's computer somewhere.

    Smiling face, because I'm self-hosting a Nextcloud server on Raspberry Pi 4.

    1. Anonymous Coward
      Anonymous Coward

      Re: Nextcloud server on Raspberry Pi 4.

      Only a single Pi 4? Tut-tut. Have you got a decent CPU heatsink as they do tend to run a tad hot at times.

      I would have hoped that you would have had a beowolf cluster of them by now!

    2. James12345
      Paris Hilton

      Re: Smiley

      Wow, all your users must be really enjoying all that storage and super fast performance.

  3. Anonymous Coward
    Anonymous Coward

    A clear case of all your eggs

    in one basket

    As has been said, we used to measure uptimes as good 99.99999%. Now? 95%

    But hey, this cloudy stuff is soooooooooooo cheap. Too bad that your business has gone TITSUP (again). What price that eh?

    Posting AC as my current employer has announced that it will soon be Google this and Google that for everything. Better start dusting off my CV as they won't survive BREXIT day as there won't be anyone left who understands what system does what at the current rate of attrition.

    1. Martin M

      Re: A clear case of all your eggs

      Which systems are you aware of that were designed for 7 nines application level availability? Did they achieve it over any kind of sensible period, e.g. years? What technologies and processes did they use to achieve less than 263ms / month downtime?

      I’ve worked on some pretty critical systems that were designed for, and achieved, 99.99% availability. They cost a flipping fortune.

      One replaced a VMS cluster, usually recognised to be a pretty reliable infrastructure, where an experienced operator made a mistake and caused a two day outage. That rather screwed its availability stats, and nearly took down a very large business.

      Very few applications have ever actually achieved even 99.999%.

      1. Anonymous Coward
        Anonymous Coward

        Re: A clear case of all your eggs

        Very few applications have ever actually achieved even 99.999%.

        Telecoms networks have 5-nines as a minimum requirement, 6-nines is not uncommon. They meet it.

        Sadly people are so used to computers & internet devices being cheap & unreliable that no-one cares any more. Until it all goes wrong, then heads must roll. Never the heads of the people who wouldn't pay for a 5-nines service, of course.

        1. DS999 Silver badge

          Re: A clear case of all your eggs

          Yep. Recently had major major storm action where my mom lives. She lost power for several days, and her DSL took another day after that before it returned. Never lost her landline.

          1. Claptrap314 Silver badge

            Re: A clear case of all your eggs

            And what is the bandwidth of that landline?

            Higher bandwidth means narrower engineering tolerances.

            And when I was a kid, the landline would go down for days at least once most years.

    2. Yougottalaugh
      Happy

      Re: A clear case of all your eggs

      Interesting conclusion to reach when Google is the only large public cloud provider that actually helps you replicate workloads in AWS and Azure. So any really critical workloads can be replicated across the big three public clouds in your choice of reasons. All your eggs in as many baskets as you want to feel safe and secure with Mother-Hen-Level engineering

    3. swm Silver badge

      Re: A clear case of all your eggs

      At Dartmouth (around 1970) we ran a 100+ user time sharing service and kept careful watch on uptime. A reboot was accomplished in less than a minute but was charged as a 15 minute interruption. A building power failure, operator error etc. were also charged with a minimum 15 minute outage. We hovered around 98-99% availability during scheduled hours. adding 9's is very hard as duplicating machinery etc. makes for more points for failure.

  4. chivo243 Silver badge
    Coat

    Fun times!

    No, it's not us... no problem here. It's on google's end. User: "Can you call google and find out how long?!"

    It's almost pub:30 getting my coat now!

  5. mr_souter_Working

    Outage or hacked?

    Google has a major outage - and I get a prompt from Chrome that some of my passwords have been breached - coincidence?

  6. Keven E

    Out again...

    ...or perhaps still.

    12 PDT

  7. martynhare

    At least it’s still news when GSuite has outages.... they must have cried tears of Azure at the thought...

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like

Biting the hand that feeds IT © 1998–2022