Amazon's New AI Fixes Server Crashes While You Sleep

Amazon Web Services (AWS) just announced a new tool that uses AI to automatically fix problems with cloud servers. When a website or app goes down, engineers often have to wake up in the middle of the night to figure out what went wrong and fix it. This new system, called AWS DevOps Agent, uses AI to investigate the issue and suggest a solution. Then, a separate automated workflow takes that suggestion and prepares the exact changes needed—but waits for a human to click 'approve' before making any changes to live systems. This keeps things safe while saving time.

The system works by using AI to look at logs, metrics, and other data to find the root cause. Once it has a diagnosis, it can automatically run read-only checks (like looking at settings) without asking. But if it needs to change something—like updating a security rule—it pauses and sends a request for approval. The engineer can then approve with one click, and the fix is applied. This reduces the time it takes to fix problems from hours to minutes, and it means engineers don't have to do repetitive detective work.

For everyday people, this means the apps and websites you rely on—like banking, shopping, or streaming—could have fewer outages and faster recovery when something goes wrong. It also means companies can save money by needing fewer engineers on call at night. The system is built on Amazon's cloud services, so it's available to businesses that use AWS. While it's not yet a fully autonomous fixer, it's a big step toward self-healing systems that keep the digital world running smoothly.