The Cluster Event Log, found on the compute (cluster) details page, is the primary tool for reviewing lifecycle events. It records every autoscale resize event (both up and down) with a precise timestamp, the reason for the event, and the system component that triggered it. This allows a Data Engineer to correlate capacity changes with specific user activities or heavy query execution times.

• Workspace audit logs track high-level administrative actions (e.g., notebook deletions, permission changes) but do not provide granular, per-cluster scaling logic.
• Driver and Executor log files contain application-level stdout and stderr for debugging code performance or failures, rather than infrastructure scaling events.
• Ganglia / Metrics UI provides a visual representation of CPU and memory usage, showing that capacity changed, but it does not provide the metadata explaining the 'why' or 'who' behind a resize trigger.