Better Error Messages for LLMs (and Everyone Else)

Vague error messages were once tolerable—but in an era of LLM-powered assistants, every message becomes part of the interface. By shifting to narrative, context-rich error reporting, you empower both humans and models to make sense of failures and act.

Silence is Golden: Filtering GitHub Notifications the Hard Way

From Push to Pull: Our Journey to GitOps with Argo CD

This year, we embarked on a journey to transition Keboola’s deployment model from a push-based to a pull-based approach, guided by GitOps principles.

Improving MySQL Performance with Table Partitioning

This is the story of how we tackled challenges with MySQL performance, why traditional solutions didn't suffice, and how we ultimately leveraged table partitioning to efficiently manage our large datasets without causing downtime.