OAuth token deletion job fails when there are multiple ASGs active

[timmc-edx]

We've been getting a number of alerts for the token deletion job failing, and most or all of these seem to be caused by the job being kicked off while multiple ASGs are active:

https://github.com/openedx/configuration/blob/70c147f6dcecfe56d370b8928b857a8cf82882b5/playbooks/active_instances_in_asg.py#L90

Possible solutions:

• Ignore the failure under these specific conditions
  • May not be feasible, as the management command isn't even getting a chance to run -- where would we add the ignore? And the alert body in OpsGenie doesn't contain the needed information.
• Change the active-instances script to use the newest ASG when there are multiple
  • Would all users of this script be OK with this?
  • Might need a new CLI param that enables this, and an opt-in for the specific job.
    • Config: https://github.com/edx/edx-internal/blob/9419ba1fa104f8be096bb69a412b2059071e6d4d/ansible/vars/prod-edx.yml#L1057
    • Jenkins job: https://github.com/edx/jenkins-job-dsl-internal/blob/837996dfa0ed0d3a5820fc745fd640a124a539bf/src/main/groovy/org/edx/jenkins/dsl/RunManagementCommands.groovy

[robrap]

Maybe we just tag these as "reviewed" automatically when closed so we get to ignore these?