11

What tools are able to provide an almost real time weather map of usage and average traffic per timeframe going over the different links?

What are some ways or solutions to actively monitor my network and provide alerts on heavy network usage over a prolonged period?

| improve this question | |
  • Everyone is mentioning Cacti with the Weathermap plugin, but no one has mentioned the Threshold plugin. Check it out, docs.cacti.net/plugin:thold That will give you your desired "High link usage alerts" and similar – jwbensley May 8 '13 at 11:14
15

There are a few examples that can do this.

Cacti is one. It has a weather-map plugin that can be leveraged to produce output such as the following

GigaPop Illinois

Wisconsin University

These are examples of what the weather map plugin can do. When configured correctly you can see minute, hourly, daily, weekly and monthly traffic stats.

You can download Cacti from the website listed above, weather-map here or a nice script hosted by Brent Salisbury here.

It is worth reading up on Cacti as there is a lot of features and devices that it can look after. Email alerts based on thresholds, events, lack of information and more can be delivered. This can be utilized via SNMP.

| improve this answer | |
  • 4
    +1 for PHP Network Weathermap. It's also worth noting that the Cacti plugin you mention above is actually PHP Network Weathermap wrapped in to Cacti. – SimonJGreen May 7 '13 at 21:55
  • 1
    -1 for Cacti; I hate having to manually specify every interface to be monitored. I don't have any better suggestions, but Cacti is much to needy for my liking. – Paul Gear May 8 '13 at 8:01
  • please add the icinga/nagios addon nagvis: docs.nagvis.org/1.5/en_US/lines_weathermap_style.html – Oluf Lorenzen May 8 '13 at 9:59
  • 1
    @PaulGear If you don't like manually adding items to monitor, and I can't believe no one has mentioned this, observium.org/wiki/Main_Page is the way forward! It auto detects interfaces and drives and temperature monitors etc, and graphs them automatically. Also as you add interfaces to a router for example, it auto graphs them automatically too! Cacti is much more power in terms of flexibility though, in Cacti I graph anything that produces a number (death rate of cats in Brazil!), Observium graphs routers, switches, firewalls, servers etc only. – jwbensley May 19 '13 at 10:48
  • 1
    I've actually written a few patches for Observium. I use it and it has great features, but it's not without its flaws. – Paul Gear May 20 '13 at 5:07
7

Wow, this question could be an entire site of its own. Regardless of what software package(s) you end up using, if snmp is how you're getting your numbers, it's time to embrace sFlow (aka: NetFlow depending on your hardware). I love snmp. It's perfect for getting a plethora of network (and other) statistics. That said, if you want a deeper look into your traffic: into vlans and protocols, start looking at whether your gear supports sFlow/Netflow.

I'm sure there are many opinions about what software to use, but here are a few that I've used in the past to put together a fairly complete monitoring solution:

  • Cacti: For graphing of everything under the sun. With additional modules, it will do some alerting on thresholds that you configure. It also supports multiple users with individual rights ... so if you're a provider and want or need to give access to customers to their own traffic graphs, this is a great tool.

  • Nagios: For monitoring and alerting. Nagios is very powerful. It has a huge arsenal of scripts, and plugins, and add-ons. It is also one of the most unintuitive platforms to configure and manage. It is a beast.

  • Zabbix: For monitoring, alerting and graphing. This is functionally cacti + nagios + quite a bit more. It will do some auto-discovery for you as well as inventory of everything it is looking at.

  • Intermapper: This is commercial software. It is monitoring and alerting. I mention it because you mention wanting something of a weather map. You can watch everything you are monitoring in real-time (or near real-time) with Intermapper. It is great software for a NOC environment.

There are many many network monitoring packages out there. These are just a few that I've had success with in the past. Good luck!

| improve this answer | |
  • What would you suggest as far as sflow/netflow software? Anything interesting in the free/cheap variety? – Luke May 7 '13 at 22:51
  • 1
    @Luke I recently started using PRTG specifically for NetFlow and have had a very positive experience so far. They also have mobile apps if you like to check the graphs/maps that way. As far as pricing it's free for 10 sensors, 1 NetFlow device = 1 sensor. – bigbash May 7 '13 at 23:56
  • Reporting back with my PRTG experience. Yea, it's windows-only, yea it's non-free, but it's slick. Point it at an IP and it spits out a very close approximation to what it is and what needs to be trended. – Luke May 10 '13 at 18:23
6

Cacti is a good open source solution. You will need some Linux experience to get up and running.

PRTG is a nice solution, can't be the price. You can use SNMP to monitor interfaces and it now also includes Netflow. Probably not the most robust reporting, but a good starting point for the price.

SolarWinds is another huge player. Great features and reporting. Only downside is you will pay a higher cost. IMO, the benefits justify the cost, but every scenario is different.

ManageEngine has some nice products, but I have always found their interface and navigation not as intuitive as other solutions on the market. Either way the product does what is advertised.

| improve this answer | |
  • You can get a ready-to-go Cacti installer from CactiEZ to run on either metal or a VM, which drastically reduces the Linux experience needed. Cacti itself will also run on Windows, with IIS. – AnotherHowie Feb 6 '14 at 11:57
4

You can use tools like Cacti or PRTG. These tools support weathermaps showing you actual bandwidth usage in the network. You can also set thresholds so that when more than 80% on a link is utilized send an e-mail or SNMP trap and so on.

These tools are mostly SNMP based. You can also use Netflow to get more visibility into what is using how much bandwidth.

| improve this answer | |
1

NfSen/nfdump (http://nfsen.sourceforge.net/) are really good for netflow. Stores your data in RRD format, and has a lot of options for displaying and processing.

| improve this answer | |
0

I have used Liveaction and it does a great job of showing a nearly real time weather map of our network. This can generate a lot of netflow and/or SNMP traffic from the end devices to the server but the information that it provides has enabled us to resolve congestion issues rather quickly.

| improve this answer | |

Not the answer you're looking for? Browse other questions tagged or ask your own question.