Olly for SREs: 3 ways I actually use it in production

A practical walkthrough of using Olly, Coralogix's autonomous AI agent, during production incident investigation. Covers three phases: identifying the primary driver by correlating metrics and logs, defining the failure boundary across service dependencies, and testing root cause hypotheses. In a real checkout error incident, Olly traced a TypeError in the checkout function to a specific commit and suggested a fix, compressing the orientation phase from ~20 minutes to ~5 minutes.