Comment # 4 on bug 1117938 from Eric Schirra

(In reply to Franck Bui from comment #3)
> (In reply to Eric Schirra from comment #2)
> > 
> > Franck i think this package can not use for send a information mail when
> > timer rise up errors.
> 
> Why not ?
> 
> Maybe there's a confusion here: systemd timers are units that are supposed
> to activate *services* on a given time.
I know it. I'm not stupid.

> So you're probably more interested in monitoring the service activated by
> the timer rather than the timer itself. And even if the timer unit itself
> fails that should be logged too.
No, i think it does not work.
systemd-denotify checks in a couple of minutes fi a service is running.
But a service for systemd.timer, perhaps, does only run for some seconds per
day.
So, when systemd-denotify checks every 30min. It comes every 30min an error
mesage. Not what i want or cron do.
Please read the doc and conf of systemd-denotify.

> And to monitor services (and everything thrown in the journal), solutions
> (more or less complex) already exist.
I know about nagios, zabbix, centreon, icinga, icinga2 and so on.
But why should i use for a new thing which does not operation without extra
complex software, when it exist a old software, which so simple effective
operation?