job timeout into permanent stuck job

Summary

using GitLab Community Edition 12.7.5 and kubernetes docker executor we get a permanent stuck job

Steps to reproduce

create a functional kubernetes docker executor on google k8s, after that scale the nodepool to 0, after that commit a change to gilab ce, the job will fail with timeout, after you restore the nodepool every subsequent pipeline will fail with stuck job, even if you remove the kubernetess executor and register a new one.

What is the current bug behavior?

every subsequent pipeline on stuck state

What is the expected correct behavior?

after the first failed job the subsequent pipelines must work normal

Relevant logs and/or screenshots