Centralize Monitoring for multiple remote sites

[shabbirali]

Is it possible to setup one single Nagios XI monitoring solution at central location may be at a datacenter to monitor 350+ host present at 4 remote sites located at different geographical locations. I have attached the diagram for better understanding.

Can we monitor all the hosts at these sites without any VPN connectivity between sites and datacenter? How the monitoring will work?
Will there be any performance issue or delay in checks or alert notifications due to latency?

What is the monitoring architect design do you recommend for such type of setup?

[SDohmen]

We have the same and we are using a distributed monitoring solution.

In our case the "offices" are our customers and each has there own nagios (core in our case) install. There we configure each service and host for that particular client. Because all customers are kinda the same we made templates for each.

On the same box we configure the NRDP client to send all data to our main central. Each customer gets there own token so if there is a security issue, it can be fixed quite fast. Below you see the commands we use for that.

Code: Select all

define command{
        command_name    submit_check_result
        command_line    /usr/bin/php /usr/local/nrdp/clients/send_nrdp.php --url=https://central/nrdp --token=sometoken --host=$HOSTNAME$ --service='$SERVICEDESC$' --state=$SERVICESTATEID$ --output='$SERVICEOUTPUT$ | $SERVICEPERFDATA$ [$SERVICECHECKCOMMAND$]'
}

define command{
        command_name    submit_host_result
        command_line    /usr/bin/php /usr/local/nrdp/clients/send_nrdp.php --url=https://central/nrdp --token=sometoken --host=$HOSTNAME$ --state=$HOSTSTATEID$ --output='$HOSTOUTPUT$ | $HOSTPERFDATA$'
}

Then we configure each host on the main central (can be done with the wizard unconfigured objects). Since the main central only recieves passive results we made seperate host templates which make sure the hosts that are attached are purely passive and have a freshness check added.

Other options for the distributed monitoring would be DNX or using NSCA although i have no experience with the first and the second is limited on the results being sent.

[shabbirali]

In our case we want only one Nagios at a central location monitoring all the sites. There will be no Nagios installed at any of the sites. We want a cloud type solution only difference will be NagiosXI will be placed at a rented datacenter so that we only need to manage one nagios across the company.

[scottwilkerson]

In our case we want only one Nagios at a central location monitoring all the sites. There will be no Nagios installed at any of the sites. We want a cloud type solution only difference will be NagiosXI will be placed at a rented datacenter so that we only need to manage one nagios across the company.

This most certainly will work just fine as long as the the Nagios XI server can reach the other networks either via public Internet or a VPN

[SDohmen]

In our case we want only one Nagios at a central location monitoring all the sites. There will be no Nagios installed at any of the sites. We want a cloud type solution only difference will be NagiosXI will be placed at a rented datacenter so that we only need to manage one nagios across the company.

If you dont want local nagios machines and you have no VPN there are only 2 options that i know off.

Option 1: Have enough open ports for each server which the central nagios can reach over internet. This is quite unsafe though.

Option 2: Use NSCA on each server. This means however that you have passive results only and means you have to configure the central to check the freshness of the data as well.

There is a option 3 which uses NRPE on all machines but use only 1 machine to send/recieve the commands from the central. I am not sure how this works though. If you want to check this out i recommend the nsclient++ package which has this ability.

[mguthrie]

Another way to do this would be to have an XI install for each location, so all of the checks can be done inside of the firewall. Then you could tie them all together either with Nagios Fusion, or you could forward the check results to a central XI server which can be used as a central viewer.