As seen in BF-34286, the aggregate() cmd is being starved by the balancer scheduling moveCollection operations because moveCollection continuously changes the collection's UUID(more details in the bf). Currently, balancer suites schedule moveCollections every ~250ms leaving little time for the agg cmd to run without being interrupted by changes to the collection metadata. Reducing the frequency of the balancer for this test will allow successful completion of the aggregation.