AI-Powered Log Analysis Automatically Diagnoses Failed Packit Builds

Breaking: Log Detective Now Integrated into Packit

PACKIT SERVICE USERS will now receive automated AI-powered analysis of failed scratch Koji builds triggered by dist-git pull requests. Starting this month, the Log Detective tool is integrated directly into Packit, eliminating the need for manual log inspection. The feature activates instantly on failure, delivering diagnostic insights within the Packit dashboard without any extra setup.

“This is a major step toward lowering the barrier for new contributors,” said Ana Martínez, lead developer of Log Detective at Red Hat. “Instead of digging through megabytes of build logs, you get a clear explanation of what went wrong and how to fix it.”

Background

Packit has long bridged upstream projects with Fedora distributions by automating builds and tests. However, diagnosing failures remained a manual, time-consuming process. Log Detective was first introduced in Copr, where users could click “Ask AI” to analyze logs. Now the same capability is embedded in Packit’s failure pipeline.

Version 4.0 of Log Detective is built on the BeeAI Framework—an agent-based architecture that processes all build artifacts. The agent employs the Drain template mining algorithm to extract concise log snippets, reducing token usage and analysis time while focusing on relevant errors.

How It Works

When a Koji build fails, Packit sends logs and metadata to the Log Detective interface server—a lightweight containerized service. The server runs the analysis and posts results to the Fedora Messaging bus, where Packit retrieves them and links them to the triggering pull request. No user configuration is required; the system automatically handles selection of logs and prompts.

“Our agent is designed to be efficient even with smaller models,” explained Martínez. “By extracting only the most informative snippets, we keep costs low and turnaround fast.” The final analysis includes a statement of the issue and—where possible—a suggested fix, based solely on the build logs.

Limitations and Target Audience

The tool is explicitly not a replacement for experienced packagers. It has no access to external knowledge bases or package-specific history. “If you’ve been maintaining packages for years, this likely won’t tell you anything new,” said Martínez. “But for newcomers, it can be a lifeline.”

Log Detective is meant to accelerate onboarding and reduce frustration for those unfamiliar with Fedora’s build ecosystem. Future versions may extend capabilities, but the current release prioritizes simplicity and immediate utility.

What This Means

For the Fedora community, this integration democratizes build debugging. Instead of relying solely on veteran maintainers, new contributors can self-serve basic diagnostics. This could increase the velocity of contributions and lower the learning curve for dist-git workflows.

“We’re not replacing human expertise—we’re amplifying it,” Martínez added. “Log Detective handles the grunt work so people can focus on the tricky problems.” The feature is live now and requires no activation; any Packit-triggered Koji build failure will automatically receive an analysis.