I built a small CLI tool to help with something that's been bugging me for a while – watching logs and trying to figure out what went wrong.

Hawkeye watches logs from files, stdin, or Docker containers. It uses simple pattern matching (regex for things like "error","exception", "404", etc.) to detect potential issues, then batches them and sends to an LLM for analysis only when something looks suspicious. The idea was to keep API costs low while still getting useful insights.

Basic usage:

heye watch docker:my-container heye watch /var/log/app.log kubectl logs -f pod-name | heye watch -

When it detects issues, it gives you a summary, possible root cause, and suggested actions. You can also ask questions about your logs with `heye ask "any database connection issues?"`. It's pretty simple and probably rough around the edges. I've been using it to monitor my current project and it's saved me some debugging time. Wanted to share in case it's useful for anyone else.

Thanks for taking a look!