# Kubernetes Local Cluster Demonstration Kubernetes demonstration on your local machine ## Prerequisites 1. Make sure to have docker or a OCI compatibile container engine installed 1. Make sure the docker/oci engine is running 1. Install `kubectl` for interacting with the kube api: - https://kubernetes.io/docs/tasks/tools/ 1. Install `kind` - mac/brew: `brew install kind` - mac/macports: `sudo port selfupdate && sudo port install kind` - win/choco: `choco install kind` - win/winget: `winget install Kubernetes.kind` - linux: tbd 1. Have at least 8GB RAM and 20GB disk ## Create cluster 1. `kind create cluster --config cluster-config/config.yaml` (1 control plane, 3 nodes) 1. Once complete, look at what is running in all namespaces: `kubectl get pods -A` ``` $ kubectl get pods -A NAMESPACE NAME READY STATUS RESTARTS AGE kube-system coredns-76f75df574-ddznd 1/1 Running 0 7m16s kube-system coredns-76f75df574-ftv2c 1/1 Running 0 7m16s kube-system etcd-demo-control-plane 1/1 Running 0 7m31s kube-system kindnet-jj8bv 1/1 Running 0 7m16s kube-system kube-apiserver-demo-control-plane 1/1 Running 0 7m31s kube-system kube-controller-manager-demo-control-plane 1/1 Running 0 7m31s kube-system kube-proxy-2wr8l 1/1 Running 0 7m16s kube-system kube-scheduler-demo-control-plane 1/1 Running 0 7m31s local-path-storage local-path-provisioner-7577fdbbfb-jj2b7 1/1 Running 0 7m16s ``` You will want to wait for all pods to become ready (showing `1/1` in the `READY` field) and once complete you will have a cluster running on your computer. ## Internal Services services that help kubernetes do what it does ### Ingress Controller (nginx) Ingress controllers manage incoming web traffic (usually HTTP/HTTPS) 1. `kubectl apply -f https://raw.githubusercontent.com/kubernetes/ingress-nginx/main/deploy/static/provider/kind/deploy.yaml` Wait for it to become available: ```bash kubectl wait --namespace ingress-nginx \ --for=condition=ready pod \ --selector=app.kubernetes.io/component=controller \ --timeout=300s ``` NOTE: You might run into problems running this on windows without WSL2. I have not tested this. ### MetalLB - LoadBalancer LoadBalancers define L2/L3 networking on the host and services as a whole define the boundary between internal k8s networking and the host's network (potentially out to internet routable ips, depending on configuration) https://kind.sigs.k8s.io/docs/user/loadbalancer/ 1. install: `kubectl apply -f https://raw.githubusercontent.com/metallb/metallb/v0.13.7/config/manifests/metallb-native.yaml` 1. list ips available to docker: `docker network inspect -f '{{.IPAM.Config}}' kind` in my case, the `172.18.0.0/16` CIDR is available: ``` [{172.18.0.0/16 172.18.0.1 map[]} {fc00:xxxx:xxxx:xxxx::/64 map[]}] ``` If yours is different, edit the [metallb/ips.yaml](./metallb/ips.yaml) file to reflect a block available to docker 1. apply ip configuration: `kubectl apply -f metallb/ips.yaml` You can tell this works when any LoadBalancer you make gets an `External-IP` ``` NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE service/grafana LoadBalancer 10.96.122.31 172.18.255.200 3000:32105/TCP 5m13s ``` ## External Services services that are run and exposed to the world ### Observability (grafana, prometheus) #### Grafana https://grafana.com/docs/grafana/latest/setup-grafana/installation/kubernetes/ 1. create namespace for deployment `kubectl create ns grafana` 1. review what we're creating in the [deployment.yaml](./grafana/deployment.yaml) 1. apply configuration `kubectl apply -n grafana -f grafana/deployment.yaml` 1. find the external ip that was given to the loadbalancer with `kubectl get service -n grafana` ``` NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE grafana LoadBalancer 10.96.122.31 172.18.255.200 3000:32105/TCP 7m43s ``` Open in a web browser or `curl` the http://:3000 to check you can access it. Default user/pass is `admin`/`admin`. #### kube-prometheus This is a pre-built subset of a ton of tools for monitoring of your cluster. It installs it's own grafana and a ton of other tools that can be used to monitor not only the cluster but everything running inside the cluster. 1. check out the [kube-prometheus](https://github.com/prometheus-operator/kube-prometheus) project: `git clone https://github.com/prometheus-operator/kube-prometheus` 1. run setup and install: ``` bash kubectl apply --server-side -f kube-prometheus/manifests/setup kubectl wait \ --for condition=Established \ --all CustomResourceDefinition \ --namespace=monitoring kubectl apply -f kube-prometheus/manifests/ ``` To [access the GUIs](https://github.com/prometheus-operator/kube-prometheus/blob/main/docs/access-ui.md), the default configuration does not provide any access outside the cluster and leaves it up to you to decide how you want to access it. For now, you can temporarily expose these outside the cluster with the following: - Prometheus: `kubectl --namespace monitoring port-forward svc/prometheus-k8s 9090` exposes http://localhost:9090 - Grafana: `kubectl --namespace monitoring port-forward svc/grafana 3000` exposes http://localhost:3000 - AlertManager: `kubectl --namespace monitoring port-forward svc/alertmanager-main 9093` exposes http://localhost:9093 The next step to exposing those services would be to use the `ingress-nginx` installation we installed above to expose the service using an nginx reverse proxy. ### Vault (secrets) tbd ### Forgejo (with redis and postgresql) tbd - https://artifacthub.io/packages/helm/forgejo-helm/forgejo#single-pod-configurations ### Sublinks in k8s tbd ## Troubleshooting tbd ## Cleanup 1. `kind delete cluster` ## Other - Rallly - https://hub.docker.com/r/lukevella/rallly https://support.rallly.co/self-hosting/introduction - Zipline - https://github.com/diced/zipline - Synapse - https://github.com/element-hq/synapse - Vaultwarden - https://github.com/dani-garcia/vaultwarden - input - https://getinput.co/ - grafana - https://grafana.com/