M/Monit should allow a method to 'gracefully' shutdown hosts

[Renny Koshy]

I'm using M/Monit in a dynamic environment where we scale up or scale down depending on traffic.

When we scale up and then down, M/Monit starts alerting about the missing servers, even though we have done it on purpose.

There needs to be a graceful way to "ignore" certain servers, or set them as "Dynamic" so that M/Monit does not complain when they go away, and instead, simply removes them from monitoring.

Maybe there needs to be some coordination between M/Monit and our systems prior to shutdown of each of these hosts.

This is critical in cloud environments.

[Tildeslash]

You just need to stop monit gracefully on the host which is to be stopped - monit then sends a "stop" event which switches the state to "stopped". If you just kill the host, M/Monit will send a heartbeat event.

If you need to cleanup hosts from M/Monit which are no longer used, you can use the HTTP-API to purge inactive entries ... see the "delete" method: https://mmonit.com/documentation/http-api/Methods/Admin_Hosts