Re: monit without start/stop

[Kimo Rosenbaum]

> The alert settings is shared by all testing rules in given  service. If you 
>want alert for one rule only, you can split the rules to two  separate checks 
>like this:


Thanks but this doesn't address what I'm seeing. I'm not using the alert 
functionality.

What I'm seeing is that every interval, if the service check fails and the 
previous state was fail, the exec script is ran and messages are logged. 
Whereas 
if the service check passes and the previous state was pass, then the exec 
script is not run and no messages are logged.

Expected behavior:

previous state -> current state | result
fail -> fail | no exec, no message
fail -> success | exec, message
success -> success | no exec, no message
success -> fail | exec, message

But the observed behavior of the fail -> fail case is 'exec, message'. 
Shouldn't 
the behavior of fail -> fail and success -> success be the same?

Thanks
Kimo