Behind the Canvas Crash

If you found yourself staring at a blank Canvas page on Tuesday, October 21st, you weren’t alone. A massive, hours-long outage of Amazon Web Services (AWS), one of the world’s largest cloud computing platforms, created a domino effect that disrupted thousands of online services globally. From popular apps like Signal, Snapchat, and Duolingo to the very core of IMSA’s academic workflow, the internet stumbled, revealing the intricate and sometimes fragile digital ecosystem we all depend on.

The root cause, as detailed by Amazon in a subsequent report, was a “latent defect” within an automated system responsible for managing the Domain Name System (DNS) for its DynamoDB database. In simpler terms, a hidden bug in Amazon’s own robotic traffic cop caused a catastrophic failure. This system is designed to constantly update and reroute internet traffic to ensure speed and reliability. However, this bug created an empty DNS record in its US-East-1 data center, and the automation meant to fix such errors itself broke down. This required manual intervention from engineers, leading to a cascading failure that left countless services, including Canvas, inaccessible.

From IMSA’s perspective, this global technical meltdown had a very local impact: a complete standstill on Canvas. After reaching out to Dr. Rowley and Dr. Glazer, they both referred me to Mr. John Chapman, the new director of ITS at IMSA. “The DynamoDB database, which routes things around, went down,” Chapman explained. “Cloud computing is subject to failure at certain times, but it’s so rare because 99.999% of the time it works.”

The most critical failure, he noted, was in the automated backup systems. In a properly functioning scenario, when one part of the AWS network fails, traffic should be instantly and seamlessly rerouted to healthy servers in another region. “For some reason, AWS didn’t route to certain regions,” Chapman said. “The DNS should be able to reroute to central or west, but it didn’t. It was a fluke incident.” He likened the challenge of preparing for such an event to “trying to prepare for a surprise tornado.”

During the outage, the ITS department’s hands were largely tied. Mr. Chapman said he “made a few phone calls, but they were at the mercy of AWS,” as the resolution depended entirely on Amazon’s engineering teams, who were scrambling to fix the core automation bug. A silver lining from the event was the validation of IMSA’s strategy of digital diversification. Because services like Google Workspace (Drive, Gmail) are hosted on a completely separate cloud infrastructure, they remained fully operational. “Only the LMS [Learning Management System] portion was affected, not Google,” Chapman confirmed. This separation prevented a total collapse of the academy’s digital tools.

In the aftermath, the ITS department is using this incident as a critical learning opportunity. “Diversifying applications will be helpful in the future,” Chapman stated, emphasizing that the department’s ongoing mission is to avoid concentrating services in a single location. “The ITS department makes sure we’re diversified across different clouds.” While a repeat of this specific, large-scale AWS failure is unlikely, the event has underscored the need for continuous evaluation of service resilience. “If this is a repeat event, other options are always something to explore,” Chapman noted, though he acknowledged the inherent unpredictability, comparing it to “getting a flat tire because you don’t know when something like this will happen.”

Ultimately, this event was more than a simple inconvenience. It was a valuable lesson in the importance of digital infrastructure because it demonstrated how a single bug in a system we may never see can ripple through our daily academic lives.

Sources:

https://www.theguardian.com/technology/2025/oct/24/amazon-reveals-cause-of-aws-outage