prairiesky
join:2008-12-08
Springstein, MB
| nagios question
Hey guys,
I've just set up nagios. It looks pretty neat, I'm going to use it to replace IP check.
I've found the code to add a ping sensor, but does anyone know of a plugin that turns it into a nice gui so it's easy to do a hundred times?
thanks! |
|
kortag
join:2009-03-11
Sandy, OR
| I have always edited the text files. It goes pretty quick. If you do end up finding a GUI let me know... but if you go with txt editing let me make a few suggestions.
1. Use host groups and service groups!
2. Break up your config files (example I have a config file for backhaul radios, customer facing AP's, switches, routers, servers, etc) It makes life way easier when you have to find a host to edit.
I'm sure my usage of nagios would be considered basic, but it serves my needs and I am pretty much always working on an issue, or already have it fixed by the time a customer realizes they are down. So I'm a big fan! |
|
anonnag
@covad.net
| reply to prairiesky
Configuration gui's are pretty much non existent except for commercial products. I have used Nagios since its Netsaint days and only recently switch to another open source product. Zenoss. so far I am fairly happy it. It even uses the Nagios plugins. All configuration is done via gui and you can make template groups. Add a system to monitor and all you checks get added via the templates. Plus it does performance graphing as well. So far it is working out for me. YMMV!! |
|
PSWired
join:2006-03-26
Edgewater, MD | reply to prairiesky
Nagat I don't think is maintained anymore, but might be what you're looking for:
»nagat.sourceforge.net/ |
|
Jerm
join:2000-04-10
Richland, WA | My work uses Xymon AKA Hobbit AKA Big Brother
Simpler than Nagios I think, but perhaps not.
»www.xymon.com/ |
|
rgoldberg
join:2008-01-05
Duluth, MN
| reply to prairiesky
I've also used nagios since the netsaint days. I can't tell you how many time I said to myself "I'm just gonna fricken write an sql driven config manager for nagios". Anyhow, in the meantime, apparently these guys did:
»www.opsview.org/
So we're in the process of moving our 1000ish hosts in to opsview.
Ryan |
|
dr mongolia
join:2008-07-03
United State
 ·Cox HSI
| Wow, Opsview looks like a significant improvement over Nagios. I especially like the way they have the host hierarchies, and of course the integration of RRDtool and historical data (including graphs) is huge. Also the API looks very easy to follow if you want to import host status into other applications, etc.
I can't believe I've never heard of this before.. Is there any reason NOT to switch to Opsview (besides migration)? I'm googling around looking for complaints/bugs but I can't find much. |
|
rgoldberg
join:2008-01-05
Duluth, MN
| No, I don't think so. No problems so far, having moved our services and the root of the ISP network into it. Migrating sucks, but we were overdue for a big clean up anyway.
Just to be clear (and to give nagios its props) nagios is the major component under the hood of opsview. The opsview guys really have done an awesome job of front-ending the config management and integrating some key plugins (basic trending).
For many years, we've run nagios (for everything) and cacti (for some things). About a year ago I started looking around at other solutions. Check out »en.wikipedia.org/wiki/Comparison···_systems. We wound up buying solarwinds orion for customer-facing reporting and are now moving to opsview and keeping cacti for our internal stuff.
The solarwinds purchase was nearly 100% driven by our high-end customers' familiarity/comfort with it. It's pretty cool too, but it's a pig (resource wise) and heavy customization, while possible, is a pain. And of course it costs money.
By the way, I don't think cacti can be beat for customized trending data, commercial or free.
Ryan |
|
dr mongolia
join:2008-07-03
United State
·Cox HSI
| Wow, thanks for that link. I'm doing some overhauling right now on our NMS so that table is great.
Yeah from the screencast I saw on their website, Opsview's RRDtool integration is not versatile enough to replace a power tool like cacti. But, IMHO, the lack of any real trending in Nagios was annoying. |
|
spectrumhead
join:2009-05-03 | reply to prairiesky
Why not The Dude by Mikrotik ?
Im using it and works for me. |
|
prairiesky
join:2008-12-08
Springstein, MB | reply to prairiesky
the dude is windows only, I'm looking for linux.
Who here has set up opsview on ubuntu? Anyone want to vnc in and set it up for me? |
|
dr mongolia
join:2008-07-03
United State
·Cox HSI
| This is unrelated to the original topic, but if you use VNC a lot you might want to look at a program called "No Machine" (they have a free version and a pay-version, there's also another free variant called FreeNX). Basically you run No Machine on your linux box and then you can remotely access it from any other computer, just like with VNC, but it runs 10X faster. With a decent connection you can't even tell that it's a remote desktop, responsiveness is just as fast as if it were local. I believe they accomplish this by SSH compression (and it's all over an SSH tunnel, which makes it much more secure than VNC). |
|
rb384997
join:2006-10-31
Athens, OH | There is a management gui for nagios, it's called Nagios QL.
It is a web application, tricky to get setup, but works well. |
|
iansltx
join:2007-02-19
Golden, CO
·Comcast
 ·Qwest.net
 ·magicjack.com
 ·BeeCreek Communica..
·Sprint Mobile Broa..
| reply to dr mongolia
I've used NoMachine and VNC on Linux stuff. While VNC isn't as secure, TightVNC on the client and server (or Screen Sharing on the client) is still fast. My beef with NX is that the memory footprint on-server is huge (100MB or so I think). Of course, if you've got the RAM then go ahead... |
|
dr mongolia
join:2008-07-03
United State
·Cox HSI
| reply to prairiesky
We've made the switch to Opsview as recommended by rgoldberg and I wanted to comment on how impressed I am with the software. It's basically everything that nagios has, plus additional ease-of-use features and automatic trending/graphing. Anyone who likes nagios should at least check it out. |
|
sizzler
@rr.com | reply to prairiesky
Has anyone got any experience with open NMS? »www.opennms.org/index.php/Main_Page |
|
viperm
Carpe Diem
Premium
join:2002-07-09
Winchester, CA
| reply to prairiesky
Are you running opsview as a internal network monitoring system I.E core routers, Ap's client radios etc?
Does it have the capabillity to page or send a text message when soemthing is down and or back up?
Ability to not send pages for 20 million devices if a parent device is down?
Ability to run ping tests from a gui to a client radio or router etc.
Thanks
--
ComTrain Certified Tower Climber. American Tower Certified approved contractor. Wireless consultants. |
|
dr mongolia
join:2008-07-03
United State
·Cox HSI
| Viperm -- In my instance it's running on an internal server and does monitor internal servers, but it also monitors remote hosts, etc. It does have the ability to text message or e-mail, and it's very flexible in this regard (certain users can be notified about certain things but not others, etc). It does also understand parent/child relationships (you set these up manually) so that you don't get hit with a million texts when a parent goes down. As far as running a test ping, it's doing this constantly, but you could schedule an additional service check and force it to immediately ping, etc. |
|
viperm
Carpe Diem
Premium
join:2002-07-09
Winchester, CA
| reply to prairiesky
Reason I wanted to do the additional "manual" ping was to have the ability for customer service agents who are NOT tech save click on a gui login and click on customer name to send a set of 4-10 pings large packets to see if the customer CPE is up or down and how well it performs etc.
--
ComTrain Certified Tower Climber. American Tower Certified approved contractor. Wireless consultants. |
|
dr mongolia
join:2008-07-03
United State
·Cox HSI
| Ah, I see what you mean. So basically you want the CPE the be monitored normally (via small packets etc), and then have the option of easily sending large pings if the customer calls in?
That exact functionality would probably be best implemented as an external shell script. There's a workaround, but it's not a manual ping: add an additional service check to your CPEs called "large ping" with a packet size of 1400 or so, then turn off notification for that service check. That way the service check would continuously run and you'd have a historical graph of the performance results, but you wouldn't get notified.
I'm beginning to consider some troubleshooting features similar to the "easy-to-use manual ping" idea that you're looking for. I'd like to have a base "troubleshooting page" where the tech enters in the MAC address of the client and it pulls up all relevant information (from monitoring system, traffic accounting, cacti graphs, etc). Ideally it would also have a manual ping option, a "traffic sniff" option, list support tickets for that particular, list all maintenance history for the AP that the user is connected to, etc. Assuming all the management systems in place have APIs this is very possible, and it would make outsourcing tech support a walk in the park. |
|
