Pointed out by a code review bot - an oversight in our design/poc for this project:

This new ifr section duplicates the existing incrementalRollout section for the same subsystem.

unknownWireFlagsDropped in particular is the same counter (getUnknownWireFlagsDroppedCount()) already reachable elsewhere and slated to land in incrementalRollout via a sibling in-flight change — exposing it a second time here under a different top-level key means the same int64 gets serialized twice per serverStatus/FTDC snapshot, and gives dashboards two names for one number.

Was folding cumulativeWireInstalls and unknownWireFlagErrors into IncrementalRolloutServerStatusSection::generateSection instead considered? That would give operators one place to look for IFR health instead of two overlapping sections.

Tracking this as a separate work item, since two in-flight PRs contribute to this new section.