Description:

When executing Python code using GraalVM, we're observing abnormal memory consumption patterns. Specifically:

The runtime creates an excessive number of threads named python-actions-pool-xxx-thread-N (where xxx is a numeric ID).
Under concurrency:
For every Python execution thread, we see a corresponding python-actions-pool-xxx thread created
At 1000 execution threads, we observe approximately 1000 python-actions-pool-* threads generated

Critical Questions:

Is the 1:1 relationship between execution threads and python-actions-pool-* threads intentional?

Does this thread proliferation directly contribute to the observed memory bloat?
What mechanisms exist to throttle/pool these threads to prevent linear scaling?

Environment:

GraalVM JDK17