AEM cloud scheduler concurrency question.

[TB3dock]

We want to create a scheduler to refresh OATH tokens which various services use.

if we set the scheduler's "concurrency" flag to true, does this mean that:

1. scheduler runce on each cloud instance.
2. scheduler runs on one instance, but concurrent with other (different) schedulers
3. scheduler runs on one instance, but can run multiple copies in parallel.

Is there any way to ensure it runs on every instance?

If any instance didnt have this running, it would mean that service on that instance would not get a new token and fail.

Is there any way to see how many instance of a cloud env are running at a given time?  We could create a scheduler to log "hello" every hour, and see how many hellos we get in the log files, but this is not very scientific.

[Jörg_Hoh · Accepted solution]

The "concurrency" flag of the Sling Scheduler just relates to the very same instance. 

Assuming that you want to start a job every 5 seconds and it normally takes just 1 second, but sometimes it takes longer. In that case the "concurrency" flag controles if the scheduler should start a new instance of the scheduled job after 5 seconds, no matter if the previous execution has been finished.