Description
Background: I'd like to filter out Kubernetes startup/liveness/readiness probes from our export PxL script. I observe that the probes always originate from the IP address of the node (host machine) on which the pod runs. If there was a way to discover the node IP from within the PxL script, then I could modify my script to exclude events with that source IP.
Describe the solution you'd like
I see there's a pod_id_to_node_name function already, so I suppose I could use a node_name_to_node_ips function.
Describe alternatives you've considered
I can't think of any alternatives to the narrow question of filtering events based on node IP.
For the broader question of filtering out k8s startup/liveness/readiness probes, the primary alternative would be to match against the http request User-Agent header, which is documented to default to kube-probe/.... However, because that default may be over-ridden and because other clients could set their own headers that way, this approach could have both false-positives and false-negatives.
Additional context
@aimichelle suggested a path forward on implementation in this slack thread.