What problem are you trying to solve?

Focus on the what and why of the need you have, not the how you'd like it solved.

We want to improve the actionability of the Host experienced a rollback alert. Currently, this alert triggers on any metadata-level rollback ID increment, including routine crash recoveries after an unclean shutdown (like cloud VM maintenance). In these cases, no data is actually reverted, so there is no action for an engineer to take. We want to avoid high-severity alerts for "events" that do not require human intervention.

What would you like to see happen?

Describe the desired outcome or enhancement.

We would like the ability to configure this alert to trigger only when a functional rollback occurs (i.e., when data is actually reverted and written to the rollback/ directory). If a rollback ID is incremented but zero Oplog entries are rolled back, the alert should either not fire or be categorizable as a lower-priority "Informational" event.

Why is this important to you or your team?

Explain how the request adds value or solves a business need.

This is important for maintaining an efficient on-call rotation. We want our high-severity alerts to be 100% actionable. When an engineer receives a "Rollback" alert, their immediate assumption is data loss. Discovering it was a false positive after manual investigation wastes engineering time and creates a "crying wolf" effect with critical monitoring.

What steps, if any, are you taking today to manage this problem?

We are currently forced to manually investigate every alert by checking the filesystem and parsing logs for "Rollback summary" strings to determine if the event was just a metadata reset or an actual data loss event.