Housing discovers a runaway Lakeflow Job rewriting an entire gold table from a bad parameter. What control should they use immediately?

When Housing discovers a runaway Lakeflow Job rewriting an entire gold table from a bad parameter, the immediate control is to stop the runaway job using Lakeflow Jobs stop controls before further damage. Stopping halts continued incorrect writes, limits cloud cost, and preserves more recoverable gold state than pausing the Lakeflow Jobs schedule using pause controls to halt the runaway update, which fails because pausing leaves the bad run actively rewriting gold. Pausing the Lakeflow Jobs schedule using pause controls to halt the runaway update fails because it leaves the bad run actively rewriting gold, so corruption and cost continue unchecked. Updating the Lakeflow Jobs run parameter using edit controls to fix the active run fails because it leaves the current bad run running with the bad value, so the damage continues while you only affect future runs. Restoring the gold table using Delta time travel controls to recover good data fails because it is a recovery step that does not stop the ongoing bad writes, and the current run continues to corrupt gold while you restore. Exam caveat: DP-750 operational incident response prioritizes stop and cancel controls on misbehaving municipal jobs before secondary actions like pausing schedules, updating run parameters, or restoring tables. Operational check: issue Stop on the runaway Lakeflow Job, verify no further gold writes, fix the parameter in source control, and repair or restore gold from a known-good version before redeploying.