feat: handle interruptions as a transient error

[rwong2888]

Motivation

I have had this in my TRANSIENT_ERROR_PATTERN to handle automatically retrying pods that have been removed for one reason or another. This is to add it to the upstream argo-workflows. I feel it would be useful for users, especially those running workflows on preemptive or spot instances.

Modifications

Added below as a transient error. (OOMKilled|imminent node shutdown|node was low on resource|pod deleted|node is shutting down|Deadline exceeded)

Verification

Used the TRANSIENT_ERROR_PATTERN at Forbes.

Documentation

This file is already referenced in the documentation.

[blkperl]

/retest

[rwong2888]

/retest

[rwong2888]

/retest