`count(std::Object)` is within our top 10 highest load queries in RDS

[jackfischer]

The nice-to-have count query is apparently very expensive in our production instance. This doesn't seem worth running in a production environment just so that we can use Gel UI.

-- {"tag": "gel/ui"}
-- {"query": "select\n    count(std::Object)", "type": 1, "extras": "{\"cc\": {\"allow_bare_ddl\": \"NeverAllow\", \"allow_dml_in_functions\": false, \"allow_user_specified_id\": false, \"apply_access_policies\": false, \"default_transaction_access_mode\": \"ReadWrite\", \"default_transaction_isolation\": \"RepeatableRead\", \"force_database_error\": \"false\", \"query_cache_mode\": \"Default\", \"simple_scoping\": null, \"store_migration_sdl\": \"NeverStore\", \"warn_old_scoping\": null}, \"pv\": [3, 0], \"of\": \"BINARY\", \"e1\": false, \"il\": 0, \"ii\": true, \"in\": true, \"io\": true, \"dn\": \"default\"}", "id": "088ebd8b-d831-6016-d8a8-86f0aba7f619", "edgeql": "select count(std::Object)"}
SELECT edgedb_v6_2f20b3fed0.__qh_76bd0e87ad71e8f09cb216ddd3900f88()