Configure service task retry to run forever with no backoff. What’s the worst that could happen? Oh right, DDoS your own API and take down production. But at least the task will eventually succeed!