High Objects Scanned / Returned constantly when a search index is added

We have a cluster that has one Atlas Search Index on it. The collection with the search index has around 5k documents. The index is an autocomplete index on one field. We were getting alerts about Scanned Objects / Returned this morning so I started looking into it. I noticed that there were large clusters of queries every 30 minutes with high scanned objects / returned, and looking into them the appName was always mongot steady state. Since this is related to Atlas Search, I deleted our search index. The scanned objects / returned metric immediately dropped to around 1, where it stayed until I added the search index back.


Is there anything I can do to help the search index be maintained without all the scanning? Should I just ignore these alerts?

I did set dynamic to false when I re-created the index, but that does not seem to have made much of a difference.

I have since learned that since the oplog is shared by all the collections, and that is what is being read to keep the index up-to-date, a lot of CRUD operations in a different collection could cause the behavior we’re seeing. Since we do have a lot of writes every half-hour to a different collection, this is likely the issue.