The project represented by SERVER-131665 guarantees predictable performance of sharding metadata refreshes regardless of the volume of expired persisted chunks on a shard. This improvement is planned to be backported to 9.0.

However, the follow-up work of reaping expired persisted chunks is likely not backportable. A large volume of expired persisted chunks may cause unintended side effects. For instance, restoring from a backup may entail updating all chunk entries' shard identifiers in a linear fashion, which may affect the RTO.

This ticket is to expose telemetry to identify shards that have accumulated a significant number of expired chunk entries.