Normally, the platform python pipeline job should collect a bunch of IP addresses (for workers) and shard the tests out onto them. You will see something like this in the Console logs:
Jeremy saw this, in which no IP addresses were collected and the test all ran on one shard:
Link to offending build: https://build.testeng.edx.org/job/edx-platform-python-pipeline-pr/9739/
This causes the job to run for very long and times out.