The query below should fail with the error that timeseries collections are not allowed with rankFusion, but it fails with a different error message. There is no validation specifically for rankFusion and timeseries in subpipelines, we just happen to fail somewhere during execution.

coll.runCommand("aggregate", {pipeline: [ {$limit: 1}, {$lookup: {from: timeseriesCollName, pipeline: rankFusionPipeline, as: "out"}}, {$unwind: "$out"}, {$replaceRoot: {newRoot: "$out"}} ], cursor: {}});

This error is happening because the validation to restrict rankFusion running on timeseries collections only exists in run_aggregate and cluster_aggregate. Subpipelines are not made nor executed using these commands if the subpipeline can be executed on the local node.