Kubernetes CNI Migration & Benchmarking
Budget: ₹1,500 – ₹12,500 INR
My production Kubernetes cluster needs hands-on assistance migrating the networking layer from its current, rather intricate CNI to either Cilium (eBPF with WireGuard) or Calico with WireGuard encryption. The overriding goal is a seamless, zero-downtime cut-over while preserving all of our existing custom NetworkPolicies—east-west, ingress, and egress traffic must continue to behave exactly as it does today.
Scope of work
The engagement starts with a deep-dive into the present topology, routing rules, and policy objects. From there, a step-by-step migration playbook and rollback plan must be crafted, executed first in staging, then in production. Once live, I need empirical data that compares both solutions: latency, throughput, CPU and memory overhead, and encryption overhead under real-world load.
Deliverables
• Migration plan with clearly defined checkpoints and rollback steps
• Fully implemented transition to either Cilium or Calico in production, no service disruption
• Verification that every custom NetworkPolicy remains effective post-migration
• Benchmark report (graphs and narrative) contrasting Cilium vs. Calico WireGuard on the cluster, using tools such as iperf, kube-burner, or similar, plus Prometheus/Grafana metrics exports
Acceptance criteria
– Cluster remains healthy and schedulable throughout the process
– All policies validated via automated tests after cut-over
– Benchmark demonstrates equal or improved performance versus the legacy setup and offers a clear recommendation
Tooling expected: kubectl, Helm, WireGuard, eBPF, iptables/nftables, Prometheus, Grafana, and version control (GitOps via ArgoCD/Terraform manifests).
Please outline a concise approach and any relevant success stories so we can move forward quickly.
Scope of work
The engagement starts with a deep-dive into the present topology, routing rules, and policy objects. From there, a step-by-step migration playbook and rollback plan must be crafted, executed first in staging, then in production. Once live, I need empirical data that compares both solutions: latency, throughput, CPU and memory overhead, and encryption overhead under real-world load.
Deliverables
• Migration plan with clearly defined checkpoints and rollback steps
• Fully implemented transition to either Cilium or Calico in production, no service disruption
• Verification that every custom NetworkPolicy remains effective post-migration
• Benchmark report (graphs and narrative) contrasting Cilium vs. Calico WireGuard on the cluster, using tools such as iperf, kube-burner, or similar, plus Prometheus/Grafana metrics exports
Acceptance criteria
– Cluster remains healthy and schedulable throughout the process
– All policies validated via automated tests after cut-over
– Benchmark demonstrates equal or improved performance versus the legacy setup and offers a clear recommendation
Tooling expected: kubectl, Helm, WireGuard, eBPF, iptables/nftables, Prometheus, Grafana, and version control (GitOps via ArgoCD/Terraform manifests).
Please outline a concise approach and any relevant success stories so we can move forward quickly.