DevOps & Infrastructure
K3s Distributions Compared: Flannel, Calico, Cilium — CNI performance benchmarks, NetworkPolicy reality (Flannel doesn't enforce them), distribution comparison (k3s official, k3sup, k3d, AutoK3s), and production deployment considerations.
Edge Computing Patterns with k3s — Multi-cluster management with Rancher Fleet (10,000+ clusters), GitOps deployments, airgap installations, private registry (Harbor), P2P registry mirror (Spegel), and observability at scale (Prometheus + Loki + Thanos/Cortex, bandwidth-efficient telemetry, bbolt for outages).
k3s in Production: Upgrades, Backup, and Disaster Recovery — Production hardening for k3s: upgrade strategies (yum/apt, blue-green, system-upgrade-controller), backup/restore (SQLite vs etcd), disaster recovery planning (RPO/RTO by business criticality, split-brain prevention, failover testing), and known issues at scale (etcd startup loop at ~2500 nodes, kubelet sandbox regression, crash on no default route, pod-to-API timeouts).
k3s Disaster Recovery Testing and Edge Patterns — Comprehensive DR testing for k3s: etcd-snapshot-schedule-cron configuration, failover testing (embedded etcd, PostgreSQL HA, Rancher Fleet multi-cluster), edge computing DR patterns (local-first backup, site-level recovery, RPO/RTO planning by network tier), Restic integration for PV backup, and validation checklist.
NetworkPolicy Enforcement: eBPF vs iptables — Cilium eBPF vs Calico iptables performance and internals. eBPF L3/L4: -0.7% overhead; L7: -87% overhead. Calico L7 needs Envoy sidecar (Enterprise). Migration strategies and debugging.
Disaster Recovery for Self-Hosted Infrastructure — Backup strategies, DR planning, and recovery validation for small-to-medium environments. Because sovereignty includes recoverability.
Production-tested guides on containerization, orchestration, observability, security, and infrastructure automation — built from real deployments.