zenping correlation

By default, zenping attempts to correlate all zenping instances on the same collector host. If zenping instances are on different physical hosts, they must share the same redis instance. In this case, the setting --redis-url which controls what redis instance zenping is talking to needs to be set to point to the same redis instance.

There are other correlation settings to use on the command line that you need to be aware of:
  • --disable-correlator: This will set the ping correlation value to False and will turn off correlation.
  • --correlation-backend: Default value is distributed, which is the default behavior when using the redis instance. If you set this to simple, no communication with redis is available and correlation will only be performed within the same zenping instance.
  • --connected-ip-suppress: Defaults to disabled. This is used when zenping does not or cannot monitor all the interfaces of routers along a traceroute path. In this case, zenping can use the model information to infer if an IP along the traceroute path belongs to a device (like a router). It will use this information to determine whether the root cause for a ping down is the "connected" device.
  • --delay-count: Defaults to 0. This means that as soon as a ping down is detected an event is sent. When using distributed ping correlation, zenping instances are not in sync. --delay-count can be set to 1 or higher in order to allow zenping instances to share the ping state before sending events. It can also be used to avoid sending events after N ping downs. For example, you may only want to send an event after 3 ping downs.