Analyze Calico to Cilium migration with minimal impact for a running cluster

Since we are analyzing the feasibility of the migration of the default CNI from open-source Calico to open-source Cilium across Sylva clusters, we need to perform a detailed analysis and work on testing/automating the steps required to achieve this migration with minimal impact to running workloads.

Starting from the work done in issue and by following the dual-overlays Cilium migration procedure, we should try to migrate a cluster from Calico to Cilium and observe the impact on running services exposed by the cluster, focusing on:

• reachability of service types ClusterIP and LoadBalancer(NodePort if possible), since iptables kube-proxy will be replaced by Cilium's eBPF implementation;
• pod-to-pod reachability (especially between Calico pods and pods already migrated to Cilium);
• impact on pods using network policies(iptables replaced with eBPF);