AWS goes down, taking Asana, Duolingo, Wordle, Amazon Prime Video and Canva with it – update

If you woke up this morning and none of your favourite web services worked, don’t worry: it’s not you, it’s them. The “them” being Amazon Web Services, AWS, which underpins so many of the websites we rely on.

“We can confirm increased error rates and latencies for multiple AWS Services in the US-EAST-1 Region,” wrote AWS this morning. “We are actively engaged and working to both mitigate the issue and understand root cause. We will provide an update in 45 minutes, or sooner if we have additional information to share.”

That message was 16 minutes ago as I write, and while it might be 1am in Pacific Time this is going to cause a huge amount of disruption in and around Europe.

This article was updated at 11.30am GMT/12.30pm UK time to reflect the changing situation. The update is a the foot of this article.

Downdetector reveals extent of problems

To see just how bad the problem is, I headed to Downdetector. This is a handy site that records the uptime, and downtime, of all the major services. You can also use Freshworks’ “Is it Down?” to check on a specific site you’re interested in.

I’ve compiled a full list of services AWS says are currently down, but check Downdetector yourself for the latest list:

What you’ll see when you try to log on

Whilst this outage continues, this is the kind of message you’ll see on affected services:

Or you might see “Maintenance break, come back later” or similar. Either way, there’s nothing you can do other than wait it out.

But it does remind us all that we do need a backup for when things go wrong out of our control. For now, I suggest you sit back and enjoy Davey Winder’s opinion piece on why disaster recovery plans should be written down on paper.

Update on services at 11.30am GMT

We are now seeing a slow recovery of services as AWS continues to work on a “full recovery”.

“The underlying DNS issue has been fully mitigated,” it wrote at 3.35am PDT (10.35am GMT), “and most AWS Service operations are succeeding normally now. Some requests may be throttled while we work toward full resolution.”

It added that “some services are continuing to work through a backlog of events”, and that there are higher than usual error rates for new EC2 instances.

All of which means that you may be lucky and your favoured service will now work fine, but others could be slower than normal or simply fail.

Downdetector continues to suggest problems with almost 50 services, as shown by a selection of nine below: note the red line, with the peaks showing the point of greatest downtime. If it’s flat, there is zero downtime.

More articles to while away the time

• British companies told to use pen and paper as serious hacks forcing them offline soar
• Microsoft has unveiled vibe working… but what is vibe working?
• Key ingredients of an AI factory (and how to have one of your own)