We have a couple dozen servers located at a hosting company site. To get to the systems we ssh from our internal network to an entry point system (a generic workstation we'll call wks1) and from there jump out to the other servers.
I want to monitor the servers and have the results collected by wks1, then have wks1 forward the results to a workstation running the web interface on our internal net.
Any thoughts, references, suggestions would be most appreciated.
Thanks,
Bill
(somewhat) remote monitoring
-
- DevOps Engineer
- Posts: 19396
- Joined: Tue Nov 15, 2011 3:11 pm
- Location: Nagios Enterprises
- Contact:
Re: (somewhat) remote monitoring
Bill,
To help us understand your network infrastructure, is wks1 a machine administered by your company or is it managed by the hosting company?
Would we be correct in assuming that the couple dozen servers at the hosting site are not directly accessible without going through wks1?
Scott
To help us understand your network infrastructure, is wks1 a machine administered by your company or is it managed by the hosting company?
Would we be correct in assuming that the couple dozen servers at the hosting site are not directly accessible without going through wks1?
Scott
Re: (somewhat) remote monitoring
Yes, wks1 is at the hosting company. It is managed and administered by the hosting company but is ours to do with as needed to manage the servers we're renting from them. Yes, the only way to get to the servers is through wks1. As we're working towards a tighter security model we're trying to avoid opening any more ports between the sites than needed for production - generally 22 to/from wks1 and a public web presence on 80 /443.scottwilkerson wrote:Bill,
To help us understand your network infrastructure, is wks1 a machine administered by your company or is it managed by the hosting company?
Would we be correct in assuming that the couple dozen servers at the hosting site are not directly accessible without going through wks1?
Scott
Re: (somewhat) remote monitoring
This is a fairly common question, to which as of yet, there isn't really a very good answer.
As far as my knowledge goes you have two options available to you:
1. This requires a lot more creative thinking on your part. Configure the servers on Nagios with a passive service (or services) and use check_dummy as the host check. Set up NSclient++ or a windows check_nsca on the workstation, if you use NSclient you need to set up the scripts that will run in the NSCA section. From here you need to get creative and it's going to really depend on what kind of checks you want to run.
2. Configure Nagios to query the workstation with NRPE and parse the hostname as an argument. Set up NSClient++ on the host and create a bunch of NRPE checks that can be executed. You are essentially using NSclient++ as a command proxy.
Hopefully one of these helps you find a succesful solution!
Ninja edit: Oh conversation happened while I was typing that... hopefully Scott has some better ideas
As far as my knowledge goes you have two options available to you:
1. This requires a lot more creative thinking on your part. Configure the servers on Nagios with a passive service (or services) and use check_dummy as the host check. Set up NSclient++ or a windows check_nsca on the workstation, if you use NSclient you need to set up the scripts that will run in the NSCA section. From here you need to get creative and it's going to really depend on what kind of checks you want to run.
2. Configure Nagios to query the workstation with NRPE and parse the hostname as an argument. Set up NSClient++ on the host and create a bunch of NRPE checks that can be executed. You are essentially using NSclient++ as a command proxy.
Hopefully one of these helps you find a succesful solution!
Ninja edit: Oh conversation happened while I was typing that... hopefully Scott has some better ideas
Re: (somewhat) remote monitoring
jsmurphy wrote:This is a fairly common question, to which as of yet, there isn't really a very good answer.
As far as my knowledge goes you have two options available to you:
1. This requires a lot more creative thinking on your part. Configure the servers on Nagios with a passive service (or services) and use check_dummy as the host check. Set up NSclient++ or a windows check_nsca on the workstation, if you use NSclient you need to set up the scripts that will run in the NSCA section. From here you need to get creative and it's going to really depend on what kind of checks you want to run.
2. Configure Nagios to query the workstation with NRPE and parse the hostname as an argument. Set up NSClient++ on the host and create a bunch of NRPE checks that can be executed. You are essentially using NSclient++ as a command proxy.
Hopefully one of these helps you find a succesful solution!
Ninja edit: Oh conversation happened while I was typing that... hopefully Scott has some better ideas
Thanks. I've done a few Nagios installs but this is the first like this. It may come down to opening ports 80/443 to wks1 and just use it as the Nagios server.
-
- DevOps Engineer
- Posts: 19396
- Joined: Tue Nov 15, 2011 3:11 pm
- Location: Nagios Enterprises
- Contact:
Re: (somewhat) remote monitoring
This is kind of what I was getting at in asking who's server wks1 was. If you are free to use the server as you wish and it is compatible with Nagiosxi as far as hardware/software, I would recommend using it as your Nagios server.wcattell wrote: Thanks. I've done a few Nagios installs but this is the first like this. It may come down to opening ports 80/443 to wks1 and just use it as the Nagios server.