Hello,

Am 18/07/2017 um 15:18 schrieb KatolaZ:
> On Tue, Jul 18, 2017 at 03:18:48PM +0200, Jaromil wrote:
>>
>> hi! love this.
>>
>> suggestion:
>>
>> can there be a threshold of 10s before signaling a down?
>> or three tries?
>>
>> also: should we open a new ml "devuan-infra" ?
>>
>
> I love this as well. It might be actually good to have something like
> nagios in place, which can handle temporary glitches automatically....
>
> My2Cents

Nagios is way more powerful, on the other hand Monit is ultra easy to
setup and can also if installed on the same host that is running the
services, perform simple actions (service restart comes to mind).
It's A classical trade-off. If we have the know-how, Nagios is probably
a better thing in the long run.

As for the threshold: it was supposed to be there since day 1, but it
wasn't. I wanted to take a look but since it wasn't too bad, I had
delayed :-D. As usual, problem was between chair and keyboard:
- There were two checks: one for HTTPS connection possible, one for SSL
about to expire (15 days).
- Turns out, I had only added the delay to the first check and from what
we've seen, the second one also triggers when HTTPS connection fails.

So, I just added the delay. That means: if a service is down but back up
again in less than 2 minutes, there won't be a notification.

As for a hypothetical devuan-infra list: if the only reason are somewhat
noisy notifications, this should fix it; if infrastructure-related
discussion is getting in the way of development, then that sounds like a
good idea.
--
Evilham