we need to enable Elasticsearch integration for one namespace.
run Elasticsearch integration and Gitlab's new indexer at scale to identify problems (unknown unknowns)
optimize index size (figure out why the index is so big and how it can be optimized). Testing is already being done using clusters spinned up locally. With a cluster in prod, engineers will be able to change indexer code and see what impact that has on the index size (using a real life dataset).
At the moment of writing we do not have self-managed ES clusters, we only use Elastic Cloud. We use it for logs on gitlab.com and indexing on ops.gitlab.net. We could use it for indexing staging.gitlab.com and gitlab.com in the exact same way as it is used for ops.gitlab.net.