I'd like to receive notification whenever a port on any of network switches reaches >90% usage. How can it be done?
this makes sense what kind of switches you use. As an example, with cisco switches each port is a single sensor. There you can monitor the network traffic on this ports. You have to set up "error fields" (a maximum of usage before PRTG is sending a message to you - cisco can be monitored with SNMP, i think this is also best practise).
If you want to do it in "percent" you can change the channels from top value to percent of max. But i would take the real values. This is more exactly.
Yes, in my case I have also each port as a single sensor. It's mostly ExtremeNetwork equipment.
The problem is that PRTG doesn't take into account the port speed value. The port speed value is displayed only during adding of a new sensor.
I have lots of 1Gbit and 10Gbit ports. So how would I get a notification once a threshold relative to port speed is reached? I don't want to go through a couple of hundred switches and set notifications for each port separately, even if it's only uplinks.
it is possible to monitor unusual sensors. This means, you can activate this setting and monitor the values. The sensor displays unnormal traffic as an orange [error]. Error is in brackets because its not an error itself. It is an information for you.
PRTG Displays every unnormal traffic for this sensors. It is to you to set up a message from this [error]. How you could do it:
How-To activate unusual sensors
- Global: PRTG-Website -> configuration -> system -> monitoring / There you can Set up the values to the unusual sensors settings
- Strict: Choose a System (ex. Cisco Switch) -> settings -> last value ( in german: "AUTOMATISCHE ANALYSE DER MONITORINGDATEN" / automatic analysis of monitoring data) / Pull the mark in the box and set up the single values
- As high as you will go, as more you can leave to the Systems under it in the tree. So if you got an Group of Switches, change the setting in the Group, not in each switch itself :).
The notifications you have to set up manually, but it is possible to do it also information in a tree, but be careful, this means every sensor in the tree is getting this.
I am aware of the monitoring of unusual sensors, but if a sensor has usually barely any traffic, say 10-100Kbps, and then this gets to a 500Kbps, this is reported as an unusual value. It doesn't help me tho as I've got a lot of sensors that behave like this.
Knowing that the port usage is almost at the limit, as in report if usage gets above certain % (of speed at which interface is running) would be extremely useful. If only a SNMP traffic sensor would have a interface speed channel reporting if an interface is on a 10/100/1000/10G setting would make this possible.
I must admit I lost you with the "settings -> last value" part.. Under device -> settings, it's the usual, device name, hostname/IP, schedule.. Can you be more specific where the "automatic analysis of monitoring data" is located?
Any thought on this?
At the moment the percentage of the total bandwidth of an interface can just be calculated with a Sensor Factory Sensor. I'll put this request on our wish list for future versions of PRTG.
The "automatic analysis of monitoring data" can be found in the settings of the device or group (not in the sensor settings).
Ok, thanks. Is there a public wishlist where users can vote/suggest/request certain features, all in one place?
Please forward feature requests to email@example.com.