[throwawaymaths]

I would have guess the builders of this would have mitigated this problem by running the python in a thread? That won't hang the VM (or cause a segfault at the 1ms boundary). It might cause OS starvation in extreme cases, but you'd have to be really extreme.

[chantepierre]

There are dedicated "dirty" schedulers for long-running NIFs that by nature cannot be preempted, to avoid interfering with BEAM process scheduling : https://www.erlang.org/doc/apps/runtime_tools/scheduler.html