We should record queries using auto_explain in default branch pipelines in order to have a canonical source of query fingerprints to compare against when running in merged results pipelines on merge requests, so that we can identify newly added queries that need review and further analysis.