While testing all the queries on capella for the optimized indexes we are creating I found a few queries in run_ingest_threads (all of the specific ones) that wrongly depended on a primary index. They were querying against the id field. Any query that is going against the id can be changed to a fetch, which is much more efficient and does not depend on a primary index.
While testing all the queries on capella for the optimized indexes we are creating I found a few queries in run_ingest_threads (all of the specific ones) that wrongly depended on a primary index. They were querying against the id field. Any query that is going against the id can be changed to a fetch, which is much more efficient and does not depend on a primary index.