Hands-on review: Paessler PRTG Network Monitoring Tool
One of our customers from New Zealand describes his experiences with PRTG
By Sean Mitchell
Every Kiwi business regardless of its size is struggling with the explosion of devices on our networks. From company-owned equipment like computers, employees' own devices and some we hadn't even thought about - like TVs and coffee machines on Wi-Fi.
Paessler PRTG is a lightweight agentless network monitoring and notification solution.
The solution seems ideally sized for the New Zealand market, with support for small organisations and scalability right up into enterprise.
You, like I did, can try the solution out without limitation for 30 days at no charge. When the trial finishes it turns into freeware with a limit of 100 network sensors.
I downloaded the free trial and installed it on a Windows 10 PC. The download was just 150MB and installation was very quick.
The solution is actually broken into two applications the core software and then a probe application. In a typical setup you'd be installing the core server and the probe app on your Windows Server, with additional probes installed elsewhere on your network. These probes could be on servers or PCs at remote branch offices, on your collocated servers in an ISP's datacentre or even just distinct parts of the network. You'd only have one instance of the core server with each of the probes reporting back to the mothership regularly.
FIGURE: Unified IT monitoring: PRTG's sunburst view gives you an easy-to-read single view of your network
One of the many advantages of this spoke and hub type of model is that in the instance a branch office network is disconnected, then the probe at that remote location keeps collecting local data and when reconnected, it uploads the backlog back to the core server, enabling both sides of the issue to continue data collection. Too often we only see one end of the issue, making diagnosis difficult.
After install you can access the solution via a Windows client or via a browser. I definitely had the best experience using Google Chrome and stuck with the browser access from here on in.
The setup process is smooth and involves a reasonably detailed checklist to best describe your network, access, privileges etc., after which it starts the network discovery phase.
In my test it took just under 15 minutes to discover every device on the network. Almost all of them luckily had a form of SNMP turned on.
It really helps to enable SNMP on as many devices as you can on the network. This turns useless IP addresses in the devices list in PRTG into useful full-featured devices with lots of info to monitor. Almost every enterprise-grade device connected to a network has some form of SNMP that can be enabled.
Many of these devices did need SNMP login information to be input into PRTG as part of the setup, although it was well worth the effort for the rich info collected.
Some of the more complicated networking devices have so much information that is available to monitor via SNMP that you can download what are called MIB files from the manufacturer. These MIB files then help PRTG to map a list of available information to monitor from the device.
Over 200 pre designed sensors are built in to ease the setup process. See the full list here.
I was shocked at how much information was available across the network. An example was a simple printer, which had an incredible stream of information available; everything from very useful information like paper and ink levels through to less useful fuser temperature information and all of which was real time.
Another example is a 24 port switch, which provided over 50 data points inside its SNMP feed for PRTG to monitor.
If you take that example and extrapolate across just a small network, you can then start to collate hundreds or even thousands of data points.
Clearly most organisations wouldn't want to monitor each network port on each switch nor every toner level in each printer. That's where Paessler's tool is really clever.
It organises your network by the probe, then by device and finally by the stream of different data available from that device. You can choose which, if any, of these are to be monitored. On one device like a server you might be monitoring many tens of channels while on a printer you might not bother monitoring it at all.
Each stream of information you monitor from a device is then called a sensor. Then Paessler actually licenses their product based upon the number of sensors in your custom-made monitoring solution.
These sensors can then be setup to notify you in a plethora of different ways from emails, txt, smartphone notifications etc. One can take it even further and create rules that if a notification isn't acknowledged in a specific amount of time then other people or systems are notified. You can easily imagine a weekend solution here.
Cleverly, when setting up the devices you can show which ones are dependent upon others. So that when a central component is having issues you don't receive hundreds of notifications from everything that's dependent upon that device.
This is really important as a flood of notifications could bury the real core issue and also encourage IT department team members to ignore notifications.
The more I look the more I discover. There are tools for designing your own visual dashboards and displaying them on smart TVs. Plus options for scheduled maintenance and adding other non-network feeds. Within the idea of adding non-network feeds to monitor, almost anything is possible. Local weather feeds could alert your datacentre engineers, overwhelming call centre stats could notify HR that more team members are needed, senior management could be notified of major sales and even highly specialised information like monitoring pupil attendance in a school.
Beyond just monitoring there numerous other actions that can be taken other than just notifications. A perfect example of which is saving power by using PRTG to turn off computers after a specific time of the day.
Paessler's solution is awesome; I highly recommend that anyone running a network needs this tool. Start small and stick with monitoring the most critical elements of your organisation first. Then only your imagination will limit what you can monitor from there.