Monitoring for a network
I'm trying to get a better handle on what's happening on my network at work. I'm currently using Nagios to monitor the big things (Server health, switch health, etc), but I'd like more granular statistics (how much data is coming and going, are any links saturated, etc). A quick glance shows that MRTG, Cricket, and RRD are good things to start playing with. I'm staring to play around with MRTG, but I'm hearing rumors of scalability issues and I don't want to get deep into MRTG if I should be putting resources elsewhere. A quick sketch of the network is a few internet connections, our router, and a few switches for networking along with a few switches for VoIP as well as our PBX. I don't want to monitor every port all the time, I instead want to monitor the uplinks of the switches to have a good idea, and look at the switch if something looks awry. My question to the group is, what do you use / recommend, and do you have anything that sticks out in your head about any of these packages (good or bad)? Thanks in advance, -- Eric Martin
We used MRTG at Verisign while I was there and it was fine... I dont get that visibility at this place so I am not sure here... On Wed, Oct 24, 2012 at 12:59 PM, Eric Martin <eric.joshua.martin@gmail.com>wrote:
I'm trying to get a better handle on what's happening on my network at work. I'm currently using Nagios to monitor the big things (Server health, switch health, etc), but I'd like more granular statistics (how much data is coming and going, are any links saturated, etc). A quick glance shows that MRTG, Cricket, and RRD are good things to start playing with. I'm staring to play around with MRTG, but I'm hearing rumors of scalability issues and I don't want to get deep into MRTG if I should be putting resources elsewhere.
A quick sketch of the network is a few internet connections, our router, and a few switches for networking along with a few switches for VoIP as well as our PBX. I don't want to monitor every port all the time, I instead want to monitor the uplinks of the switches to have a good idea, and look at the switch if something looks awry.
My question to the group is, what do you use / recommend, and do you have anything that sticks out in your head about any of these packages (good or bad)?
Thanks in advance, -- Eric Martin
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
I'd take Cacti over MRTG, more flexible and scale-able IMO. Mike also mentioned OpenNMS which is pretty decent too. You could get into more involved setups with Zabbix, ZenOSS, Hyperic but the ROI on those could turn out to be minimal. I've used MRTG, Cricket and Cacti in Production and at home and prefer Cacti above them all, as it has involved the least head banging over the years. I've tried the various others but they fulfill a bigger role (combined network, system, and application monitoring) and are probably overkill for what you are trying to accomplish. Dave Date: Wed, 24 Oct 2012 12:59:26 -0400 From: eric.joshua.martin@gmail.com To: wlug@mail.wlug.org Subject: [Wlug] Monitoring for a network I'm trying to get a better handle on what's happening on my network at work. I'm currently using Nagios to monitor the big things (Server health, switch health, etc), but I'd like more granular statistics (how much data is coming and going, are any links saturated, etc). A quick glance shows that MRTG, Cricket, and RRD are good things to start playing with. I'm staring to play around with MRTG, but I'm hearing rumors of scalability issues and I don't want to get deep into MRTG if I should be putting resources elsewhere. A quick sketch of the network is a few internet connections, our router, and a few switches for networking along with a few switches for VoIP as well as our PBX. I don't want to monitor every port all the time, I instead want to monitor the uplinks of the switches to have a good idea, and look at the switch if something looks awry. My question to the group is, what do you use / recommend, and do you have anything that sticks out in your head about any of these packages (good or bad)? Thanks in advance, -- Eric Martin _______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
Cacti is cool once you get ot configured. We've been playong with graphite and we've ditched our cacti graphs in favor of it. Creating a new graph consists of openong a tcp port ane passing ina tuple... Tada... New graph. Then you can go to the web interface and tinker with the graph parameters. Just my .02. Tim. On 10/24/12, David Coutu <cout@alum.wpi.edu> wrote:
I'd take Cacti over MRTG, more flexible and scale-able IMO. Mike also mentioned OpenNMS which is pretty decent too. You could get into more involved setups with Zabbix, ZenOSS, Hyperic but the ROI on those could turn out to be minimal. I've used MRTG, Cricket and Cacti in Production and at home and prefer Cacti above them all, as it has involved the least head banging over the years. I've tried the various others but they fulfill a bigger role (combined network, system, and application monitoring) and are probably overkill for what you are trying to accomplish. Dave
Date: Wed, 24 Oct 2012 12:59:26 -0400 From: eric.joshua.martin@gmail.com To: wlug@mail.wlug.org Subject: [Wlug] Monitoring for a network
I'm trying to get a better handle on what's happening on my network at work. I'm currently using Nagios to monitor the big things (Server health, switch health, etc), but I'd like more granular statistics (how much data is coming and going, are any links saturated, etc). A quick glance shows that MRTG, Cricket, and RRD are good things to start playing with. I'm staring to play around with MRTG, but I'm hearing rumors of scalability issues and I don't want to get deep into MRTG if I should be putting resources elsewhere.
A quick sketch of the network is a few internet connections, our router, and a few switches for networking along with a few switches for VoIP as well as our PBX. I don't want to monitor every port all the time, I instead want to monitor the uplinks of the switches to have a good idea, and look at the switch if something looks awry.
My question to the group is, what do you use / recommend, and do you have anything that sticks out in your head about any of these packages (good or bad)? Thanks in advance,
-- Eric Martin
_______________________________________________ Wlug mailing list Wlug@mail.wlug.org http://mail.wlug.org/mailman/listinfo/wlug
-- Sent from my mobile device I am leery of the allegiances of any politician who refers to their constituents as "consumers".
Have not checked out Graphite yet, thanks for the heads up. Looks like it uses time series data, which is interesting. I've poked around OpenTSDB before and found the potential for usefuleness. Just haven't had the time to really dig in with it at all. Will definitely check out Graphite, as I agree that Cacti requires some manual interaction to get new hosts added and components monitored. It's one of it's major drawbacks as it definitely doesn't lend itself to auto-magical configuration when adding new hosts like Nagios/Icinga can. It can be done if you are determined, but it's not out of the box. Thanks Tim! Dave
Date: Wed, 24 Oct 2012 14:19:30 -0400 From: turbofx@gmail.com To: wlug@mail.wlug.org Subject: Re: [Wlug] Monitoring for a network
Cacti is cool once you get ot configured. We've been playong with graphite and we've ditched our cacti graphs in favor of it.
Creating a new graph consists of openong a tcp port ane passing ina tuple... Tada... New graph.
Then you can go to the web interface and tinker with the graph parameters.
Just my .02.
Tim. o/wlug
On Wed, Oct 24, 2012 at 12:59:26PM -0400, Eric Martin wrote:
I'm trying to get a better handle on what's happening on my network at work. I'm currently using Nagios to monitor the big things (Server health, switch health, etc), but I'd like more granular statistics (how much data is coming and going, are any links saturated, etc). A quick glance shows that MRTG, Cricket, and RRD are good things to start playing with. I'm staring to play around with MRTG, but I'm hearing rumors of scalability issues and I don't want to get deep into MRTG if I should be putting resources elsewhere.
A quick sketch of the network is a few internet connections, our router, and a few switches for networking along with a few switches for VoIP as well as our PBX. I don't want to monitor every port all the time, I instead want to monitor the uplinks of the switches to have a good idea, and look at the switch if something looks awry.
My question to the group is, what do you use / recommend, and do you have anything that sticks out in your head about any of these packages (good or bad)?
Statseeker is good if you want to monitor every interface @ 1 minute intervals without any rollup of the data over time that RRD does. But it is commercial. It is a FreeBSD appliance install (they basically provide an ISO, you provide the server).
participants (6)
-
Chuck Anderson -
David Coutu -
Eric Martin -
Mike Peckar -
Steve Pelland -
Tim Keller