Replit’s new agent takes your on-call shift

What changed

Replit’s Monitoring Agent is pitched as on-call for app builders. Outages don’t first reach you through frustrated users.

What it makes possible

• Alerts that come with a diagnosis. Automatic checks watch your published app’s error logs, HTTP errors and request latency. With Auto-fix on, the agent uses your app’s production metrics, runtime logs and code to find the cause, then attempts a fix and validates it.
• You decide how far it goes. Monitoring alerts, Auto-fix and Auto-republish are separate switches. Auto-fix alone doesn’t publish anything to your live app. Auto-republish is experimental and needs Auto-fix on.
• Questions without waiting for an alert. You can ask the agent about production anytime, with prompts like “Why did my app return 500 errors in the last hour?”

A possible use

Say a small booking app starts throwing errors overnight. With alerts and Auto-fix on, the agent can dig through the logs and code and prepare a validated fix, and you review it in Agent chat in the morning and republish when you’re happy.

Before you switch it on

• Not every alert ends in a fix. A problem in an outside service, for example, may need action outside your app.
• Replit’s docs warn that some changes, including migrations and index operations, may be risky, so think twice before letting it publish on its own.
• Runs may add Agent usage charges.

To set it up, open the Publishing tool, open your app’s publish settings and switch on Monitoring alerts, then Auto-fix. Each published app is configured separately.

For the longer walkthrough of the three switches and the limits, read our Main Course on Replit’s Monitoring Agent.

Want another agent that digs into alerts and ships a checked fix? See our Daily on Copilot’s CodeQL autofix.