Is it more efficient with temporal to schedule a new cron workflow after a delay, or to keep an existing workflow running and use activity retries, if the delay is moderate (10m), and thousands of workflows are running?
Activity retry is much more efficient.