Our performance tests typically run on small clusters (replicaset/1-3 shards). This leaves a gap in our performance testing for conditions that typically only occur at enterprise scale (e.g., many chunks/collections/indexes). However, actually running performance tests at the scale of our larger customers is unrealistic.

Therefore, this ticket is to investigate which potential edge cases in the catalog and routing components only occur at scale and might potentially affect performance/availability, and open tickets to write performance tests to mimic these conditions in smaller clusters (e.g., by artificially creating 1 million chunks).