[INUG-Users] SSM agent heartbeat failures

[Kiser, Matthew]

We have our ssm agents set up to send a heartbeat event once per minute,
and an automation to watch for this event and alarm if the heartbeat is
not received for 5 minutes.

 

We have a few agents which is causing us problems (currently only
observed from windows agents).  They are multihomed, and sometimes send
out the heartbeat trap from an interface other that the primary,
occasionally causing "false" heartbeat alarm.

 

Has anyone had experience with this in the past? 

Can the agent be locked down to send traps out only a specific
interface?

Is this even a function of the ssm, or would you expect it to be a
function of the systems administration team?