When performing a Process Exhaustion attack against a given pod, depending on the provisioned resources for a container it may run out of memory and the pod's main process will be oom-killed.
When a process exhaustion attack occurs, the gremlin process has an adjusted oom_score to prevent it being oom-killed, as this can create a state where the zombie processes created during a given experiment are not properly cleaned up. It is preferable for the primary container process to be oom-killed than for the gremlin process get oom-killed and leave hanging processes behind that will ultimately deprive the container of resources.
Was this article helpful?
That’s Great!
Thank you for your feedback
Sorry! We couldn't be helpful
Thank you for your feedback
Feedback sent
We appreciate your effort and will try to fix the article