Re: Feature Request: put alerts in message queues

[Jan-Henrik Haukeland]

Integrating Monit with a message bus/queue sounds like a good idea. 
Unfortunately there are a plethora of options and selecting one and integrating 
with just one is probably not a good idea. For Monit's part using a more 
abstract or general integration solution makes better sense. Luckily this is 
already in place [1], namely using the exec statement to call out to any 
program which can do the delivery. PagerDuty for instance has an example here 
http://www.pagerduty.com/docs/guides/monit-integration-guide/ Mind, not a 
message queue, but the approach should be similar.

1. That said, alert in Monit could be improved and is something we think about. 
Some discussion has been done here 
https://bitbucket.org/tildeslash/monit/pull-request/19/support-for-alerting-by-shell-commands/diff
 and we're also planning to include nanomsg in Monit. Nanomsg is kind of a 
predecessor to ZeroMQ. 


> On 23 Jan 2015, at 18:53, Steven Truong <address@hidden> wrote:
> 
> Hi all,
> 
> I have used monit for a long time and I love it.  Thanks for all the hard 
> works.
> 
> It came to my realization that email alerts are getting old and there are 
> certainly needs to tight monit into other monitoring solutions such as 
> nagios/Icinga/Sensu.  The idea is instead of getting alert emails from Monit 
> and other monitoring solution I would like to see monit submit 
> statuses/alerts to message queues such as 0MQ or the likes and then my 
> monitoring solution to read of the queues and take care of the logics and 
> decide how to alert me.  Besides that, I think there are some overlapping in 
> using monit and other monitoring solutions.
> 
> Currently, I use monit to monitor and auto restart critical services and then 
> alert.  I then also use another monitoring solution to monitor whether these 
> critical services are running or not.  
> 
> I would rather want to use monit to take care the works of monitoring and 
> restarting these critical services and submit results to queues and the other 
> monitoring solution will just handle other monitoring needs but is 
> responsible for all monitoring alerts.
> 
> There are real needs to be able to monitor and auto restart critical services 
> but there are quite a lot of overlaps in using these solutions together.
> 
> Please point me to any information that I might have missed and share your 
> thoughts on this feature request.
> 
> Thank you very much,
> Steven.