Pular para o conteúdo principal

kubernetes

DockNimbus vs Kubernetes

DockNimbus sits above Kubernetes — it can deploy K3s clusters as one of its capabilities, but it's a broader platform:

DockNimbusKubernetes
What it isSelf-hosted cloud platformContainer orchestration engine
ScopeFull stack: nodes, networking, storage, clusters, workloadsOrchestrates containers within an existing cluster
Assumes you haveBare metal machines with SSH accessAn already-provisioned cluster
Node provisioningBuilt-in — SSH installs agent, Docker, WireGuardNot included — needs kubeadm, Kubespray, Rancher, etc.
Cluster creationnimbus kubernetes create or manifest YAMLManual bootstrapping (kubeadm init, join)
Networking setupWireGuard mesh VPN auto-configured between nodesRequires CNI plugin (Calico, Flannel, Cilium)
Load balancingBuilt-in EasyHAProxy with domain routingRequires Ingress controller (nginx, Traefik) + external LB
DNSBuilt-in dnsmasq + Cloudflare integrationCoreDNS (cluster-internal only), external DNS separate
StorageNFS volumes + MinIO S3 provisioned via CLI/manifestRequires CSI drivers, PV/PVC configuration, external storage
Non-K8s workloadsDocker Swarm services, Compose stacks, standalone containersKubernetes only
IaCSingle manifest declares nodes → clusters → workloadsSeparate tools needed: Terraform (infra) + Helm/Kustomize (workloads)
StateCentralized SQLite (single source of truth for everything)etcd (cluster state only, not infra)
Config complexity~20 line YAML for a full stackDeployment + Service + Ingress + PVC + ConfigMap + ...
Multi-clusterNative — manage multiple Swarms and K8s clusters from one CLINeeds federation or tools like Rancher, Loft, Admiralty
MonitoringBasic heartbeat + hardware metricsRequires Prometheus/Grafana stack
SecuritymTLS (agents), HMAC/JWT (users), WireGuard (transit)RBAC, NetworkPolicy, PSA — but cluster must already exist
Learning curveHoursWeeks to months
Minimum resourcesSingle RPi, ~256MB RAM2GB+ RAM per node (control plane needs more)
EcosystemOpinionated, self-containedMassive — Helm charts, operators, CRDs, service meshes
ExtensibilityLimited to what DockNimbus providesNearly unlimited via CRDs, operators, admission webhooks

When to use which

DockNimbus — You want to go from bare metal to running workloads without stitching together 5+ tools. You don't need the full Kubernetes ecosystem and its extensibility. You want one CLI and one manifest format for everything from node setup to workload deployment.

Kubernetes — You need the full ecosystem: custom operators, CRDs, advanced scheduling (affinity, taints, topology), service mesh, GitOps pipelines, or you're deploying to an environment where K8s is already running (EKS, GKE, AKS). You have the team and time to manage the complexity.

Key philosophical difference

  • Kubernetes answers: "I have a cluster — how do I orchestrate containers on it?"
  • DockNimbus answers: "I have machines — how do I turn them into a working platform?"

Kubernetes is one layer of the stack. DockNimbus is the whole stack — and it happens to use K3s (or Docker Swarm) as its orchestration layer under the hood.