Despite Downdetector report, no evidence of another AWS outage

So far, the number of Downdetector reports is sitting at around 500, which is relatively low. For a major outage affecting multiple regions, for example, we'd expect a few thousand reports at least.

It's worth noting that Slack is also currently suffering an outage and it's not impossible the two incidents are connected. When AWS went down late last year, it took a whole host of sites and services with it, Slack included.

Last time round, the AWS outage was caused by a loss of power in a single data center facility in the US, which caused a host of knock-on problems.

Amazon was quick to restore power to the affected facility, but faced network connectivity challenges that delayed the recovery.

This time, it appears the outage is on a smaller scale. There's still nothing on the AWS status page and very little chatter on Twitter, including from official Amazon accounts.

Since there's nothing on the website or Twitter feeds, we've sent a message to our contacts at AWS asking for further clarification. This may be a small outage, but it's clear something is going on.

We're still in the dark as to precisely what's going on, but it's worth noting that Downdetector is showing a drop-off in the number of reports over the last few minutes.

Does this suggest someone has pulled a few levers and tweaked a few dials to set the issues straight?

Looks like some people are using the AWS and Slack outages as the perfect excuse for a quick powernap...

Although the official status page still suggests AWS services are operating as normal, the outage appears to be causing problems across the US, with issues concentrated on the east coast in New York and Boston and in Denver, Colorado.

There are also whispers that the ongoing AWS outage may be causing problems with services from exercise platform Peloton, although this is yet to be confirmed.

It looks like the AWS issues are actually getting worse, not better. The number of reports registered on Downdetector has spiked again.

Meanwhile, our AWS contact has promised to investigate and come back to us with more information soon.

In the absence of any concrete information from AWS, we want to hear from you. Pop us a message over Twitter, letting us know where you're located and the precise services affected.

The latest updates from Downdetector don't paint a pretty picture, with the number of reports now climbing to almost 800. Perhaps we were wrong to suggest this is just a small-scale outage.

We're still awaiting any kind of communication from AWS, though, which is very unusual. Amazon is usually quick to update the status page. What is going on?

Now two hours since the outage began, AWS is still yet to make a statement about the issue or update its status page, pictured below.

The plot thickens...In an email exchange with TechRadar Pro, Amazon has denied there are any issues with its service whatsoever, despite the surge in reports on Downdetector and the chatter on social media.

We have asked the company for further clarification, given the conflicting information.

The number of reports on Downdetector is now down significantly on its peak, currently sitting around the 300 mark. There's clearly something still going on, though.

A slow recovery
AWS operations appear to be returning to normal, though Downdetector doesn't indicate that we're out of the woods yet.

Amazon's AWS status page is, however, telling a very different story, offering no outage reports throughout the globe. Instead, it's all green checks and "Service is operating normally."

Who to believe
Amazon's status reports have remained clean as a whistle throughout Tuesday, but Downdetector still has an above average number of reports for the service, which would indicate there were some issues, somewhere.

Right now, AWS chatter on Twitter is at, essentially, zero, so whatever outages existed, they appear to have been resolved.

Amazon has told us it believes that reports on TechRadar Pro of an AWS service outage are incorrect. A company spokesperson provided us with the following statement:

"On February 22, 2022, there was no broad service issue with AWS. Our services were operating normally. The only resource on the internet that provides accurate data on the availability of our services is the AWS Service Health Dashboard."

We've been doing some digging around what may have caused the discrepancy between the reporting on Downdetector and the official word coming out of AWS.

Ookla, the company that owns Downdetector, told us that the widespread outage at Slack (a known AWS customer) may have been misdiagnosed by some users as a problem on the end of the cloud provider.

However, the firm also noted that the nature of the platform means Downdetector is often better at noticing small-scale problems that affect only a small pool of users, which may otherwise be missed by internal monitoring.

Here's what Luke Deryckx, Ookla CTO, had to say:

"Downdetector provides a view of service availability from the perspective of the users of that platform, using data reported by users themselves. We often find that minor problems impacting a small group of users will go unnoticed on internal monitoring systems, and Downdetector can help point the way to the source of the problem."

"We see a small number of credible reports that some AWS users had issues with their platform yesterday. However, we do not believe there was a widespread service issue on AWS’s platform. To put it in context, Downdetector received about 8,000 total user reports of problems at AWS yesterday, whereas we received over 400,000 user reports during the most recent significant AWS outage on December 16."

"Downdector independently verified service issues at AWS during the December event and published a 'widespread outage' alert. While we don’t discount the reports of those experiencing problems with their AWS-hosted services yesterday, we saw no evidence to justify a widespread AWS-related outage notice for this event."