What triggers the Scheduler to post MACHINE_UNAVAILABLE alarms like the following...
[06/04/2020 17:40:46] CAUAJM_I_40245 EVENT: ALARM ALARM: MACHINE_UNAVAILABLE MACHINE: agent_host
[06/04/2020 17:48:48] CAUAJM_W_40253 Machine <agent_host> is not responding. Taking offline.
[06/04/2020 17:48:52] CAUAJM_I_40245 EVENT: ALARM ALARM: MACHINE_UNAVAILABLE MACHINE: agent_host
[06/04/2020 17:48:52] <Machine <agent_host> did not respond to repeated ping requests.>
Release : 11.3.6
Component : CA Workload Automation AE (AutoSys)
That sequence of log messages is triggered when the scheduler tries to open a connection to an agent to run a job or as part of the agent inventory scan and the agent does not respond. When the first attempt is made that fails, the machine will be placed in an unqualified status. The scheduler will then try to ping the machine a few more times. If it doesn't succeed after those attempts, the machine is marked offline and the alarm is generated. If the agent connectivity is restored later, the scheduler will pick that up during one of the agent inventory scans and put the machine back online automatically. An agent restart will also automatically trigger the scheduler to mark it online if the connectivity is restored at that point.