This article applies to PRTG Network Monitor 15 or later
Getting Alerts for a Down Internet Link in Case of Multiple Links (Failovers)
There's no direct answer to this. It will heavily differ based on the available devices and how the connections and links are implemented. PRTG's flexibility allows you to handle various different scenarios with different approaches:
Method 1: Device's Interface Status
For starters, it is rather simple to monitor the device's interface physical connection status with SNMP, using the standard SNMP Traffic Sensor. Add this sensor to your device and configure the Connection Status Handling within the sensor's Traffic Specific settings to Show alarm for all "disconnected" states.
Method 2: Current Traffic on the Device's Interface
If the primary interface's traffic within your SNMP Traffic Sensor is constantly high (for traffic IN or OUT), you may create channel limits to warn you if traffic goes below a specific value which could indicate a failure.
This approach is very simple (minimal setup effort) but is also error prone because you may experience intervals during the day with no traffic due to irregular utilization patterns. This will also not work to monitor fail-over links because they will most likely have little traffic when in stand-by.
Method 3: Private Device's SNMP Implementation
Some devices provide extended SNMP support which may allow you to query properties like failover route status based on the results of the device's built-in reachability tests. This will depend on vendor support and should be possible to deploy with an SNMP Custom Sensor together with lookups. Please check with the vendor of your device if it supports this functionality.
Method 4: Ping from the PRTG Cloud
With the introduction of our new Cloud-based sensors a new option to monitor these scenarios was revealed. It is now possible to ping both WAN interfaces of your router/firewall from the internet. Within PRTG create two Cloud Ping sensors to ping both interfaces from the PRTG Cloud. If a sensor goes down, you will know that your device's IP for that WAN link is no longer reachable from the internet.
The solution is easy to implement and reliable. The downsides are that your router or firewall may not be allowed to be pinged from the internet and the minimum scanning interval for the Cloud sensors is currently 10 minutes.
Method 5: Monitoring Directly from PRTG Using ICMP Together with Routing Rules in Your Router/Firewall
This is the most complex solution to the problem, but it can also provide a very fast response time.
It requires you to choose two remote (internet) hosts that you ping. Then set up two static routes (no auto-failover) for PRTG to reach these locations over each distinct link.
Once you deployed both ping sensors, each of them will test an individual link. If one of them stops working (for example, goes down), it will either mean two things:
- The remote device is down.
- Your ISP's link is down.
You may need to implement additional sensors or rely on the results from the traffic sensors to confirm whenever the remote host is down or if one of your links is now down (and you are in failover).
More
You do not need to stick with a single option, you can use one or two or all of them at once. Once set up, we recommend that you also create notifications to make sure you get an alert as soon as such an event happens. It is also important to make sure that PRTG is able to deliver the notifications when one of the ISP links is down. For more details about setting up notifications, please see the following manual sections:
Add comment