Hot exploration investigates a reproducible failure while it is happening. Start where the error first appears, follow its trail through available evidence, and move progressively closer to the root cause. section 10.3.2
Use hot exploration when a failure can be triggered reliably:
- Confirm the exact steps, inputs, user actions, permissions, configuration, external systems, and datasets that trigger the problem. section 10.3.2
- Watch the failure unfold and gather evidence at the relevant moment, including logs, console output, network traces, monitoring data, recent code changes, and dependency availability. section 10.3.2
- Follow the evidence across system boundaries until it identifies the faulty operation or code location. section 10.3.2
Tracing a broken web form
Reproduce the form failure, then inspect the browser console and network trace. An HTTP 500 response from a web service shifts the investigation to the backend. Matching that timestamp against application logs reveals a stack trace and pinpoints the faulty line of code. section 10.3.2
Hot exploration depends on reproducibility. When the failure cannot be reproduced, switch to cold exploration: build a timeline from the first report and reconstruct events from remaining traces. section 10.3.2 section 10.4
