Cohort checklists we reuse before every SQL lab opens

Before participants touch a shared warehouse, we snapshot schemas and pin dependency versions in a single markdown file mentors can diff week to week. That sounds bureaucratic, but it prevents the classic failure mode where two desks solve the same bug with conflicting hotfixes.

We also rehearse a “broken pipe” minute: intentionally drop a staging view, then walk the room through the error message, the lineage note, and the fix. People remember the triage pattern better than another slide on best practices.

Finally, we assign a rotating “signal keeper” who watches queue depth and disk usage on lab machines. If you are teaching remotely across time zones, that small role cuts panic threads in half.

None of this replaces curiosity; it just keeps curiosity pointed at SQL problems instead of environment mysteries.