Sharding tasks on enterprise-rhel8-arm64-debug-tsan-all-feature-flags intermittently hang until the Evergreen task times out. The underlying issue is host memory exhaustion: TSAN is unable to mmap its trace/shadow pages across much of the fixture, producing errors such as:

{{ERROR: ThreadSanitizer: out of memory: failed to allocate 0x40000
(262144) bytes of TracePart (error code: 12)
ERROR: Failed to mmap}}

resmoke_jobs_factor is already set to 0.1, the most aggressive value used by evergreen_resmoke_job_count.py, and these sharding tasks run on a 32-vCPU 8xlarge host, so round(32 * 0.1) still yields 3 concurrent jobs, or roughly 36 TSAN mongo processes. 

Setting resmoke_jobs_max: 1 will force this variant to run a single fixture regardless of host core count.