Currently the overwrite param is set to true for all collections, except for clustered collections. The effect of this is allowing overwrites on recordIds in the record store. For example, you can insert rid(1) -> Y when some rid(1) -> X already exists and overwrite the existing mapping.

Setting it to false for all collections would mean that we would get an error if we attempted to do this. This would be a nice guardrail to have, because as far as I can see, no collection, clustered or not, should ever attempt to insert with a recordId that already exists in the record store.