There are several queries that are unlikely to finish, ever, on analytics wikireplicas:
While those servers are there to go long-running queries, there is no reason to run them if they are not going to finish ever but they create issues for other users- the longer they go on, the slower the servers become for other users:
labsdb1010 4134749 s51591 dbproxy1010 wikidatawiki_p 2d
SELECT ips_item_id, count(*) AS cnt FROM wb_items_per_site WHERE CONCAT("Q", ips_item_id) IN (SELECT page_title FROM page JOIN pagelinks ON page_id=pl_from WHERE pl_title="Q937857") AND ips_item_id NOT IN (SELECT ips_item_id FROM wb_items_per_site WHERE ips_site_id="enwiki") GROUP BY ips_item_id ORDER BY cnt DESC LIMIT 100
labsdb1010 4729760 s51591 dbproxy1010 wikidatawiki_p 1d
SELECT page_title FROM page WHERE page_namespace = 0 AND page_title NOT IN (SELECT term_full_entity_id FROM wb_terms) AND page_title NOT IN (SELECT CONCAT("Q", ips_item_id) FROM wb_items_per_site) AND page_is_redirect = 0
labsdb1010 5067887 s51591 dbproxy1010 wikidatawiki_p 6h
SELECT DISTINCT a.term_entity_id, b.term_entity_id FROM wb_terms AS a JOIN wb_terms AS b ON a.term_text=b.term_text WHERE a.term_type="label" AND b.term_type="label" AND a.term_entity_type="item" AND b.term_entity_type="item" AND a.term_entity_id <> b.term_entity_id AND a.term_language = "en" AND b.term_language <> "en" AND a.term_full_entity_id IN (SELECT page_title FROM page JOIN pagelinks ON page_id = pl_from WHERE pl_title="Q4167410") AND b.term_entity_id IN (SELECT page_title FROM page JOIN pagelinks ON page_id = pl_from WHERE pl_title="Q4167410") LIMIT 2000You should use EXPLAIN on those queries to understand why that is happening and how they could be made in smaller chunks so they can finish. Those selects without where seem highly inefficient.
See also T114904: Migrate wb_items_per_site to using prefixed entity IDs instead of numeric IDs