I have to implement an Agent. The Agent should run once per hour. Is it possible to avoid the next start if the agent activity is still activ? My problem is, I can' guarantee that processing ends within one hour.
It may not be that easy to prevent the next start, but it should be fairly easy to have the next start abort if another incarnation of the agent is still running. For example, have the Agent (or a wrapper on the Agent) set a DSS setting to "true" when running, and have it set to "false" when about to exit. If the Agent starts and sees "true", it knows not to do anything, but to immediately exit, since another incarnation of the Agent is still running.
This, or something similar, should not be too difficult to implement.