Nagios Support Forum

Hi.

I'm new to Nagios.
From what I have seen Nagios works in a pulling mode, once in a while goes to a service, checks for it and sets a status accordingly (ok,criticical etc ...).
Is there a more simple way to push random alarms to Nagios without the need to define complex scripts/various services to check ?
Any alarming system such as HP OpenView / CA / even Windows event viewer can receive many alarms from various sources and just show them.
I do not wish to add a complex service for each alarm that I want to shoot from my new developing system.
I know that there are Passive Alarms/Services and I already checked them. But again, even in this case, an event goes to a "specific" service which can be in a single state of "Ok,Critical,Warning etc ..."
Is there a solution for this in Nagios ? Like to show a dashboard of alarms with their history and push random alarms without defining complex scripts/services ?

Thank you.

In short the answer is no.

Passive checks can receive check results at random times however there still needs to be a service defined for those results to be seen in Nagios, otherwise Nagios will log an error in the log saying it received a check result for an unknown service.

If you are just doing development, then you could just have one service defined called "Passive Development Service" and then you would only ever need to submit check results to one service.

Nagios could actually exist with just one service called "the service" and you could submit all random check results to it. If the service was volatile then they would send you an email each time. However you'll never have any real service history or performance data, you're basically setting nagios as something that receives xxx and sends it off in an email.

Can you explain more why you don't want to setup services for what you want to monitor?

Thanks for your prompt answer !
Yes, the reason for me not to want to define a service for each check is complexity.
For example, I want to add 5 new alarms to my new system (let's say R&D wise).
Then, I have to coordinate with our DevOps team that a new alarm (in this case a new service) came on, and that they need to define it in the Nagios system.
This makes me and our team think 100 times before issuing such a procedure. Instead if I could just send an Alarm and keep the history of that in Nagios, we (R&D team) could
develop hundreds of different alarms in different severity levels and just submit them to Nagios. Then, I could in Nagios do a filter by "severity" or "string" etc, issue emails/SMSs whatever ....
Many systems work that way and to me this sort of behavior looks very intuitive.

Sounds like Nagios Logserver might be what you're looking for then. With Logserver you don't really define hosts and services, just ports to listen on and message formats to accept. From then on you can send just about any text-based data you would like and LS will happily archive it and allow searching.