This repo is a collection of POCs built primarily on GCP and Kubernetes. The goal is to continuously learn new tech, experiemt with new combinations, reverse engineer while also documenting and sharing the knowledge with my team and the community.
POCs that use k8s assume pre-existing cluster and access with sufficient permission to GCP and/or AWS account. For creating Kubernetes clusters I use other repos:
https://github.com/olga-mir/k8s-multi-cluster - AWS multi cluster setup provisioned with Cluster API and FluxCD https://github.com/olga-mir/k8s - AWS with kOps, GKE with terraform or gcloud (with misc flavors: ASM, multi network for pods, etc)
Most folders contain self-containted POCs or experiments. These folders are prefixed with 2-digit number in such a way that latest experiments appear at the top.
Other general purpose folders:
k8s-manifests-general - manifests to deploy on a k8s cluster
scripts - generic scripts
Scripts and files in this repo rely heavily on information provided via env vars. For GCP values file may look like this:
# Common
export PROJECT_ID=
export REGION=
export NETWORK=
export SUBNETWORK=
# Specific projects
# user mail is required for Cloud Run scenarios where invoker check is enabled
export USER_EMAIL=
Some templates also use dynamic variables and envsubst to render a manifest that can be deployed.
envsubst < alpine-targeted-node-tmpl.yaml > alpine-targeted-node-rendered.yaml
To render all tmpl found in the repo run scripts/render.sh from any location in repo dir and apply file that ends with -rendered.yaml
Login to ECR and create docker login secret in $NAMESPACE (if not set or provided in command line default test is used):
./scripts/login-private-registry.sh [NAMEPACE]