The PRTG cluster functionality provides high-availability for your monitoring environment.
A PRTG cluster consists of one primary master node, and one or more failover nodes. If the primary master node fails, one of the failover nodes will take over the role of master and will control the cluster until the master returns. This ensures fail-safe operation with no gaps in the real-time monitoring and alerting. During normal operation, all configuration is performed on the primary master node, and this configuration is automatically distributed to all failover nodes in real time. You can view data on any of the cluster nodes, but the failover nodes are in read-only mode. When you configure a cluster, PRTG automatically creates a new probe, called the cluster probe. You then create new devices in the cluster probe, or move existing devices into the cluster probe.
The PRTG cluster is an active-active cluster. Each node in the cluster will actively monitor all devices in the cluster probe at the same time. Please note that this active-active cluster generates additional monitoring traffic in your network and more monitoring load on the target devices in the cluster probe. Each node in the cluster maintains its own, separate database, so that the storage of monitoring results is distributed among the cluster. If you are using remote probes in your environment, you can configure the cluster connectivity for the remote probes. You can decide whether your remote probes connect to only the primary master node, or if they connect to all failover nodes as well.
We now have three different types of probes available: the local probe, the cluster probe, and remote probes. Each probe type has a very different method of operation. For devices and sensors on the local probe: only the primary master server will send out monitoring queries, and only the master server will store the results in its database. If the master server fails, then there is no monitoring and no alerting for the sensors on the local probe until the master server comes back online. There will be a gap in the historical data for the time that the master server was offline. For devices and sensors on the cluster probe: all of the nodes in the cluster will send out monitoring queries for these sensors, and all of the nodes in the cluster will store the results in their separate databases. If the master server fails, the failover nodes will continue to send out queries and receive the results, and will send out notifications for the sensors on the cluster probe. There is no gap in either the real-time monitoring or in the alerting, and historical data is available in the databases on the other PRTG cluster nodes. For devices and sensors on a remote probe: the one machine where the remote probe is installed is the only machine that sends out monitoring queries for those sensors. The remote probe then sends the results in real-time to either one or all of the cluster nodes, depending on your configuration. The cluster nodes then store the results in their separate databases. If the remote probe is configured to send data only to the primary master server, and the master server fails, then there is no alerting for those sensors during the time that the master server is offline. The remote probe will buffer its results and will send them to the master server when the master server comes back online. If the buffer overflows, there will be gaps in the historical data for that remote probe. If the remote probe is configured to send data to all of the cluster nodes, and the master server fails, then the remaining failover nodes will still receive data from the remote probe and will send out alerts. The failover nodes will also save the monitoring results in their separate databases. The remote probe will buffer results and will send them to the master server when the master server comes back online. If the buffer overflows, then historical data is still available in the databases on the other PRTG cluster nodes. Each PRTG cluster can have one local probe, one cluster probe, and multiple remote probes, so you can tailor the cluster and levels of redundancy to your needs. The PRTG cluster: full high-availability for your monitoring environment.
