The Amazon web hosting service has been having issues for 6 hours now and I’m kinda surprised I haven’t seen a thread here on it unless I missed it.
Our district has required us to use one curriculum service (uploading assignments, students submit assignments, grades entered, etc.). At the end of last year they cancelled teacher access to the alternatives to save money. So of course student cannot access classes or assignments. And those that can, we cannot see it on the teacher side to grade but it does look like it is save somewhere on their servers.
Incident Title: Amazon Web Services (AWS) Global Outage
Incident Number: INC0038212
Priority Level: P1
Summary: A major vendor operational issue is currently affecting multiple AWS services globally. Currently we are aware of the following issues but not exhaustive:
Online Transactions
Adding memberships to cart
Renewals
Viewing profile data
APPC portal
Testing in lower environments
Status: Investigation and mitigating activities continue.
Incident Start Time: 20 Oct 2025, 03:11 EDT
Incident Reported Time: 20 Oct 2025, 03:51 EDT
Impacted Areas: Global
Impact Statement: Users may still experience intermittent degraded performance, failed requests, or inability to access certain services.
I’m in Europe. Our public platforms are hosted on AWS infrastructure resident in Ireland.
Because the core issues are localized in the US, all of our systems have continued to function. We haven’t had any kind of outage or performance impact.
However, on the back end, many of the monitors seem to think the associated services aren’t working. “Oh no, I can’t replicate the backups!” (We check the backups, they’re fine.) “Oh no, we’ve lost your certificate registry!” (We check the certs, they’re fine.)
It’s an interesting insight into how the infrastructure is actually interdependent. None of our data goes to the US for legal reasons, and all our services run locally, but there’s some oversight component that relies on something in the US.
As a professor, I’ve had to postpone an online discussion assignment, and email a PDF to students because they can’t access it on our learning management system (Canvas, specifically).
KDP (Amazon publishing platform)'s Help page is headed
We’re currently experiencing technical issues and apologize for the disruption to some Amazon sites and services. Our teams are working to fix this as quickly as possible. We will provide updates here once the issue is resolved. Thank you for your patience.
We got an email from our IT well after the AWS went down and hours after school started explaining the issue and assuring us the website we needed (explained above) was working … it’s still not working.
I had to check on the squid. In the dark. And the remaining hamsters. None of the automation was working. No soothing background noise was playing. It was a bad morning.
IRL however, none of my amazon devices were working, which was a pain for a few of the lights, but they’re in non-critical areas. And I had to use my backup options for morning workout music (stream from my local Plex server and on-device timer). But those came back up (responding more slowly though) about an hour ago.
That really points out how much of most commercial or governmental websites have dependencies they don’t recognize. Which dependencies have further dependencies they not only don’t recognize, but can’t recognize.
It’s a very, very deep and very tangled stack of unknown unknowns.
This is illustrative, even though it’s about units of software, not groups of servers:
from xkcd: Dependency
You don’t have to pull out a very large slice of the total before the whole thing (or sections of the whole thing) come tumbling down.
