I would like to use Remote Probes with my PRTG cluster setup. How do I use Remote Probes in a cluster? Can I connect them to any node? What if the node with the Remote Probe connected fails?
This article applies to PRTG Network Monitor 8 or later
Connect Remote Probes to the Master Node
You can use Remote Probes in a PRTG cluster. These monitor the (remote) network they're installed in and forward monitoring data to the PRTG Master Node core server.
However, there are some restrictions when using Remote Probes in a cluster setup:
- You can configure Remote Probes for a connection with the Master Node only. Due to the individual authentication method, a Remote Probe is bound to the Master Node and cannot be setup for connection with a Failover Node.
- If the Master Node of a cluster fails, another Failover Node takes the Master role as long as the original Master Node is down. During this time, a Remote Probe cannot send any data, as the Failover Nodes are unknown to it. So, the probe will buffer monitoring data (this possible for a certain time) and send it once the original Master Node server is back online.
Remote Probes are a powerful monitoring feature. However, some restrictions apply when the connection to the (Master Node) core server is lost. For more information about this, please see the article Monitoring Client Computers Offline Using Remote Probes.
This is a feature that some of us have been asking for since PRTG7. I, for one, had hoped that PRTG9 would include the ability to connect remote probes to a failover server. I really don't understand why it's not possible to do this. Can you advise please when this feature will be available?
"Clustered Remote Probes" are considered, but I'm afraid I cannot give you an exact release-date/version in the moment. Please bear with us.
Please don't "Consider" but add it to the to-do list, we need this as well. When maintenance is done to the primary server you not only loose sight during that time, you also get inaccuracy in monthly reports.
Need this function as well. Please add to to-do-list
I am curious how hard this would be to implement. Assuming I am using a domain name for my remote probes to connect to as long as the DNS redirects to the failover site when the primary site fails all of the probes would be able to redirect to the failover site. As long as the failover site can accept the access key it is using it seems to like it would work. Maybe I am missing something that makes this much harder to program for.
unfortunately there is more to it. If it was easy we would have already implemented it.
Ba assured that all your votes are counted and that we consider the effort.
Having the ability to redirect the remote probes to another node that is acting as the current Master node would make PRTG more excellent than it currently is.
Clustering really has no functional value without that ability.
Is there any timeline or consideration for the addition of this feature?
Also, I do not recall, but once the communication failure that isolated the original Master node is resolved, does the current Master then update the original Master node and return control ?
This is a huge feature I am looking for as well. If I am running a clustered PRTG setup it would be natural to think that if my Master goes down the devices then re-connect to the Failover as I have the DNS set to roll to the Failover server.
all votes for this feature are counted.
+1 for this feature
Any news about this feature? I am looking forward to have cluster probe to take over as failover when the master node goes down.
The feature is still on the wishlist but no release date has been determined yet.
+1 for this feature.
I would assume that all the probes have a list of the different cluster servers. And when the master can't be reached for a period of time it asks one of the other servers it knows, what server is the new master. Just my 2 cents.
Love the product, please continue the great support!
+1 for this feature
Any news regarding this feature that seems "indefinitely postponed" ?
As mentioned earlier: if this was easy to implement, we would have already added this feature. However, unfortunately it is not. There is still no specific timeframe for when such a feature will be included - we still have it on our future implementation list, but it is unlikely that such a function will be available any time soon. Sorry.
Nothing worth doing is easy. ;)
Honestly, this is the only reason why we can't rely on PRTG for up/down monitoring of critical services. We use remote probes heavily and can't have the master server become a single point of failure in our monitoring infrastructure.