Description of problem:

Using some Kubernetes platforms with Cilium as the CNI will cause the nodes to have lots of container interfaces with lxc as a prefix. If these nodes are registered to Foreman/Satellite, the facts for interfaces will become very large and cause slowdowns. Add interfaces with "lxc*" pattern to list of IGNORED_INTERFACES.

I have created an issue and submitted a PR in upstream Foreman:

https://projects.theforeman.org/issues/38036
https://github.com/theforeman/foreman/pull/10386

How reproducible:

Run a Kubernetes node with Cilium CNI and register it to foreman/satellte. Observe large amount of reported interfaces in host facts ( hundreds )

Business Impact / Additional info:

I am a Red Hat consultant on an engagement and we have a user who has encoutnered this problem running SUSE Rancher RKE2 ( Kubernetes Cluster ) with Cilium as the CNI on RHEL 8. Seems using RHEL is supported by SUSE for running RKE2, so there's potential other customers could run into this issue using off-the-shelf software with default config and registering to Satellite.