Akin to the idea behind SERVER-31670 of reducing the number of unexpected elections observed in our tests, I'd like to propose that we use

• a 3-node CSRS where all nodes have votes=1 and priority=1 when the test suite attempts to step down the CSRS primary, and
• a 1-node CSRS for all other cases.

This would mean the following test suites would run with a 1-node CSRS

• aggregation_sharded_collections_passthrough.yml
• causally_consistent_jscore_passthrough.yml
• causally_consistent_jscore_passthrough_auth.yml
• change_streams_mongos_passthrough.yml
• change_streams_sharded_collections_passthrough.yml
• change_streams_secondary_reads.yml
• integration_tests_sharded.yml
• jstestfuzz_sharded.yml
• jstestfuzz_sharded_causal_consistency.yml
• jstestfuzz_sharded_session.yml
• sharded_causally_consistent_jscore_passthrough.yml
• sharded_collections_jscore_passthrough.yml
• sharding_gle_auth_basics_passthrough.yml
• sharding_jscore_op_query_passthrough.yml
• sharding_jscore_passthrough.yml

and only jstestfuzz_sharded_continuous_stepdown.yml would continue to use a 3-node replica set. Is the Sharding team agreeable to this proposal? I'd also be curious to know if this is something we'd consider doing for sharding-related tests that aren't interested in specifically how collection metadata operates are processed.