Verify Users should get a choice. Many failures follow resource exhaustion. Disk exhaustion and filesystem corruption present clearly in logs and by I/O errors; free or expand storage, run filesystem checks, and if the node store is corrupted consider restoring from a known good snapshot rather than attempting ad hoc repairs. These practices minimize downtime and reduce the need for...