Do I need a Backup Collector?
Backup collectors can eliminate the collector as a single point of failure, ensuring monitoring continues should the original collector go down. If a collector is declared down and a backup collector is configured, then all devices monitored by the down collector will failover to the backup collector (the collector down alert will indicate that devices monitored by the down collector were re-assigned to the backup collector). The backup collector will collect and report data so that monitoring continues.
Because the backup collector will take over all the monitoring, you must make sure that the backup collector has the same ability to collect data as the initial collector. For example, you must list both collectors in any firewalls restricting traffic to the monitored hosts and both collectors must be permitted in any specific snmpd.conf, ntp.conf, or other configurations that may limit monitoring access.
For example, you may want to consider using a backup collector if you are regularly performing maintenance on your collector machine.
We recommend that you set backup collectors in pairs (e.g. collector A is backed up by collector B and collector B is backed up by collector A).
Note that a collector defined as a backup collector can also be assigned a set of devices to collect data from (i.e. it can be active for one set of devices, and backup for another).
Setting up a backup collector
- Install or identify a second collector on a different server that can monitor the same set of devices
- Click manage for the original collector and select the backup collector in the 'Failover Collector' field
Once you select a failover collector, you'll have the additional option of enabling automatic failback:
If checked, this option will transfer the devices back to the preferred collector once it becomes available again. If this option is not checked, you can manually re-assign devices back to their preferred collector or to a new collector.