#GCP #GKE #kubernetes #gitops #argocd #gitlab #helm

Today I wanted to share with you a project that has kept me busy throughout the Christmas holidays. I have been migrating all my websites, which were previously scattered across several servers in GCP and AWS to one single cluster using Google Kubernetes Engine.
My criteria for this project was to reduce costs as much as possible, automate absolutely everything, and to serve as a training ground for experimentation with new best practices and new technologies especially Gitops with Gitlab, Terraform and ArgoCD, Helm templates for Kubernetes, and some cost-saving options that are available in the Google Kubernetes Engine, Preemptible and Spot VMs.
I had to make a few tough choices to save money. But at the same time, I didn't want to become too dependent on the cloud provider, in this case, Google Cloud Platform. Kubernetes helps a great deal in that sense since it is already cloud-agnostic. Every choice I have made in this project was with the intention of keeping costs low but without sacrificing security, incurring any data loss or performance penalty. Also I made sure I automated every single step of my setup, in case I ever need to move quickly from one cloud provider to the other.

Article:
https://spltech.co.uk/how-i-migrated-...

Timestamps:

0:00 Introduction
1:21 Github or Gitlab?
1:58 Git Large File Storage (LFS)
2:37 Gitlab Project Structure
4:00 ArgoCD
4:44 Semantic Versioning
5:32 Terraform Setup
6:25 Gitlab Integration with Kubernetes Deprecation
7:36 Running a database in Kubernetes - Mysql Operators
9:00 Developing my own Mysql database solution for Kubernetes
9:43 Load Balancing and ingress controllers
10:08 Analysing the costs for load balancing in GKE
10:46 Kubernetes Nginx Ingress Controller
12:02 Using Preemptible or Spot VMs
12:30 Disadvantages of Preemptible Nodes
13:00 Spot VMs
13:57 GKE Cluster architecture
15:00 Using Node affinity rules
15:40 A brief mention about monitoring