What Really Happens When You Upload a File to S3?
You click upload. A second later, your file is "in S3." But what actually happened?
- Your file is split and copied
S3 does not keep your file in just one place. It stores copies across multiple data centers (Availability Zones) in the same Region. If one data center fails, your file is still safe. - 11 nines of durability
S3 promises 99.999999999% durability. That means if you store 10 million objects, you might lose one every 10,000 years or so. Your file is very safe. - You get the file back right away
Since 2020, S3 has strong read-after-write consistency. Upload a file, read it a moment later, and you get the new version. No more waiting or retrying. - It's not really folders
S3 has no real folders. photos/2026/cat.jpg is just one long name (a key). The / only looks like a folder in the console. - Upload events can trigger code
This is where it gets fun. A new upload can fire an event that starts a Lambda function. You can resize images, scan files, or start a data pipeline, all without running a server.
A simple idea to try this week
Build this small flow:
User uploads an image to S3
S3 event triggers Lambda
Lambda makes a thumbnail
Thumbnail is saved to another bucket
Use two buckets, not one. If Lambda writes back to the same bucket, it can trigger itself again and again, and your bill will grow fast.
S3 looks simple, but a lot of smart engineering sits behind that upload button. Understanding it will help you build safer and cheaper apps on AWS.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article