|
The last Amazon Prime Day 2024 (July 17-18) was Amazon’s biggest Prime Day shopping event ever, with record-breaking sales and more items sold in two days than any previous Prime Day event. Prime members shopped millions of deals and saved billions of dollars across more than 35 categories worldwide.
I live in Korea, but fortunately, I was in Seattle to attend the AWS Heroes Summit during Prime Day 2024. I signed up for Prime membership and used Rufus, a new AI-based conversational shopping assistant, to quickly and easily find items. Prime members in the US, like me, consolidated shipping for millions of orders during Prime Day, saving about 10 million trips. This consolidation reduces carbon emissions on average.
Jeff’s annual blog posts show how AWS powers the Amazon website and mobile app to make these short, large-scale global events possible. (Check out his posts from 2016, 2017, 2019, 2020, 2021, 2022, and 2023 to get a little more in-depth.) Today, I want to share some of the top AWS numbers that made my amazing shopping experience possible.
Prime Day 2024 – All the numbers
Some of the most interesting and/or surprising indicators are:
Amazon EC2 – Amazon deployed a cluster of over 80,000 Inferentia and Trainium chips for Prime Day, as many Amazon.com services, like Rufus and Search, use AWS artificial intelligence (AI) chips under the hood. During Prime Day 2024, Amazon used over 250,000 AWS Graviton chips to power over 5,800 unique Amazon.com services (double the number from 2023).
Amazon EBS – Amazon provisioned 264 PiB of Amazon EBS storage in 2024 to support Prime Day, a 62% increase compared to 2023. Compared to the day before Prime Day 2024, Amazon.com’s Amazon EBS performance surged to 5.6 trillion read/write I/O operations during the event, a 64% increase compared to Prime Day 2023. Additionally, compared to the day before Prime Day 2024, Amazon.com grew and transferred 444 petabytes of data during the event, an 81% increase compared to Prime Day 2023.
Amazon Aurora – On Prime Day, 6,311 database instances running the PostgreSQL-compatible and MySQL-compatible editions of Amazon Aurora processed more than 376 billion transactions, stored 2,978 terabytes of data, and transferred 913 terabytes of data.
Amazon DynamoDB – DynamoDB powers several high-traffic Amazon properties and systems, including Alexa, Amazon.com sites, and all Amazon fulfillment centers. During Prime Day, these sources made tens of trillions of calls to DynamoDB APIs. DynamoDB provided single-digit millisecond responses and maintained high availability even as the peak reached 146 million requests per second.
Amazon ElasticCache – ElastiCache has handled over a trillion requests per day and has reached peak requests per minute.
Amazon QuickSite – During Prime Day 2024, one Amazon QuickSight dashboard used by the Prime Day team had 107,000 unique views, over 1,300 unique visitors, and delivered over 1.6 million queries.
Amazon Sagemaker – SageMaker served over 145 billion inference requests during Prime Day.
Amazon Simple Email Service (Amazon SES) – SES sent 30% more emails to Amazon.com during Prime Day 2024 compared to 2023, and delivered 99.23% of those emails to customers.
Amazon Guard Duty – During Prime Day 2024, Amazon GuardDuty monitored approximately 6 trillion log events per hour, a 31.9% increase over the previous year’s Prime Day.
AWS Cloud Trail – CloudTrail processed over 976 billion events to support Prime Day 2024.
Amazon CloudFront – CloudFront served a peak load of over 500 million HTTP requests per minute during Prime Day 2024, with over 1.3 trillion total HTTP requests, a 30% increase in total requests compared to Prime Day 2023.
Ready for expansion
As Jeff mentions every year, rigorous preparation is critical to the success of Prime Day and other large-scale events. For example, we tested resiliency with 733 AWS Fault Injection Service experiments to ensure Amazon.com was highly available on Prime Day.
If you are preparing for a similar business-critical event, product launch, or migration, we highly recommend leveraging the newly branded AWS Countdown, a support program designed for the project lifecycle to assess operational readiness, identify and mitigate risks, and plan capacity using proven playbooks developed by AWS experts. For example, with the additional help of AWS Countdown, Legal Zoom successfully migrated 450 servers with minimal issues, and continues to leverage AWS Countdown Premium to streamline and accelerate the launch of SaaS applications.
I’m looking forward to seeing what records will be broken next year!
