your suggested way is:
1- collect all documents to index in an array
2- bulk-delete all currently indexed documents from search db
3- bulk-index new documents from scratch into search db

meanwhile, this cron works slightly differently, where:
1- doesn't collect models, but instead upsert them on the spot to db
2- doesn't delete all models, but only outdated ones (runId is different)
3- doesn't have a third step

So, ... for the search implementation to follow along with this cron implementation it should:
1- index (upsert) each document on the spot, don't save it to an array and do that later.
2- bulk-delete all documents that are outdated (runId is different)

I added comments on how to do that

thank you a lot @omdxp ! one step forward to having search in the website