Nagios Support Forum

Hi and thanks for a great monitoring program but i am still a beginner i feel.
I have a nagios4 core setup running on a box with ubuntu 16.04 server which after connection to a Gateway router then a firewall it goes out to several remote vpn endpoint 3G router/modems. Some of these have devices as children. Children show unreachable when the 3G modem is down but the service monitoring the 3G modem is where i have some understanding problems.
I understand parents relating to hosts but am not sure about getting service parents to be useful or if they are needed.

Do I need both a host check as well as a service check on the 3G modem IP or should it just be only a host check?
Initially I left the host check as default values and defined a service check that was an ICMP 1000 byte packet to test fully the 3G link and performance throughput. A single ping might work ok yet there would be poor thoughput with larger data.

crude diagram.
10.0.1.100------>10.0.1.254------------------- >10.0.100.1>
ubuntu+nagios4>networkgateway vpn- to--->3G router->host check ping 10.0.100.1
------------------------------------------------------------------>service check icmp 10.0.100.1 1000 bytes.
I have the host check ping to the 3G modem as well as a service check icmp to the router as the same IP. is this doubling up unnecessary?
Now if the 3G device is down, I rely more on the icmp large packets for consideration which is a service check.
However if the 3G device fails then the service is critical and not unreachable state.

Should i put a service parent as the host check?
thanks for any ideas in working this out.

Eventually i would like to some add snmp query of the modem performance to aid in monitoring these, like signal and ecio levels but this might be more advanced for me at this point. All are Cisco 3g devices c881g or similar.

Minsik Australia.

Generally speaking, you would not need to set up service parents in such a way.

You may want to take a look at https://support.nagios.com/forum/viewto ... 10#p103689 for some additional info. Let us know if you have additional questions. I don't want to bog you down with details if if you are just looking to catch the majority of cases.

thanks for the reply however i am still unclear of the differences and when to reply on one or the other.

Can this explanation help to picture the question?
I have the need to monitor remote 3G modems and their availabillity.
Initially setting up Nagios i added the devices IP address as hosts, and also added 1 service which was the same device and IP.
Then to stress the connection a bit more than a basic ping discovered the icpm_check command so amusing this as the service check (even though the name is ping). This sends 2 icmp packet of 1000 bytes to stress test the 3G connection and prove if itd working to be able to pass some reasonable load of data.

So the original host_check is still as default with the added icmp_check as the service.

My question is should i just have the host check and try to cnfigure that as the icmp_check 1000bytes check?
It feels like the host_check and service_check is a double up of what i am trying to do.However in the future i want to add services like snmp polling for the signal level and ECIO levels and do some graphing of these values.

thanks for ideas and suggestions. All appreciated by a raw beginner.

MInsik

Yes, you could just have your host check use check_icmp and pass the -b 1000.

It's entirely up to you, but changing the check-host-alive command in the commands.cfg would change it for all of your hosts.
You can work around this however you would like.

Correct, It wouldn't make sense if you were going to do a check_icmp for both a host and service check.

As for the SNMP polling, those can just be separate services under this host.

thanks for the suggestions and help. Seems like the more experience the better.

Was there anything else (related) that you needed? Or are we all set to close this up?