Clustered Environment

Hello Michele,
thank you for your inquiry.

Please allow me to expose some figures about sensors/performance and our recommendations:

• Recommended Setup for Most PRTG Users

I also advise you to review the following in depth articles:

• PRTG Manual: System Requirements(Detailed)
• KB: How can I speed up PRTG—especially for large installations?
• I run PRTG on VMware. How can I obtain best performance?

While it may work in some conditions, we don't advise that PRTG is run on virtual deployments that exceed 2,500 sensors and consider a physical server mandatory for deployments that exceed 5,000 sensors. When working with a PRTG Cluster (with 2 nodes), these values should be divided by the number of nodes (2).

We're also not able to provide support for performance issues on deployments above 10,000 sensors (standalone) and 5,000 sensors in a 2-node cluster even when deployed on physical servers.

You're talking about a 15,000 deployment which exceeds these figures by 50% and is clustered, if you're encountering performance issues these are most likely the limits of what PRTG can currently support. The articles above provide a couple of details/recommendations to improve performance, but if you intend to reach 30,000 sensors we strongly recommend that you deploy multiple parallel core servers to balance this load. Please note that additional licenses will be required for each standalone/2-node cluster deployment.

As for the "Probe Service" that runs on your failover nodes (for the "Failover/Local" Probe), that can't be removed and we advise against disabling it, since the Probe Device contains the "Core/Probe Health" sensors which are used to measure the health of your deployment, please bear in mind that when only monitoring the probe device these probes have very little influence on the overall system performance, the guides above provide some recommendations on how to gain additional performance when working with these sensor amounts.

Feel free to contact us via a support ticket as it will give us more insights about your deployment performance, we may be able to provide additional suggestions/recommendations after reviewing a support bundle, but a "definitive solution" may not be possible with this sensor amount in a Clustered deployment.

Best Regards,
Luciano Lingnau [Paessler Support]

Hello Luciano,

thanks for the reply. The Support limitations are well known at our side, unfortunately we have this and make the best out of it. The only bottleneck we have currently is the PRTG Internal Performance itself.

CPU, RAM, Diskspace does not matter. we have 16 CPUs and 16GB Ram on both Core Machines but there is no issue at all with these ressources.

Just Config Changes and reboots of Machines takes quite a bit long

As you mentioned, we removed every Sensor of from the Core Probes. So they just check the Core Installation itself. (this made things already better)

Do you think you could find a few more tweaks if we send you a Support Bundle?

Kind Regards Michele

Support Bundle should be available for you now,
I have referred this KB Article

Hello Michele,

we've received your bundle, case PAEXXXX49.. We'll analyze it and reply shortly.

Best Regards,
Luciano Lingnau [Paessler Support]