Deploy a scalable SQL database cluster

As a brief introduction this project creates a cluster of PostgreSql servers with Streaming Replication and failover enabled.
It creates a Primary PostgreSql pod and two PostgreSql Replica pods and replicates primary's database in real-time to Replica pods.
Everything was done using the kubegres operator that really simplifies the process.
For the ease of use when running the project everything is handled by terraform.

Demo

challenge demo video

Prerequisites

create terraform.tfvars with the following variables:

do_token="digital ocean token"
pg_user="postgres user"
pg_pwd="postgres password"

Description

Before creating the k8s cluster a digital ocean project and vpc are created for grouping everything together.

resource "digitalocean_project" "do-challenge-proj" {
  name        = "do-challenge-proj"
  description = "A project for digital ocean challenge."
  purpose     = "k8s cluster"
  environment = "Development"

  resources = [
    "do:kubernetes:${digitalocean_kubernetes_cluster.doks.id}"
  ]
}

resource "digitalocean_vpc" "do-challenge-vpc" {
  name     = "do-challenge-vpc"
  region   = "lon1"
  ip_range = "10.0.0.0/24"
}

The digitalocean managed control plane cluster is created and also the autoscaled worker pool node with a min of 1 node and a max of 3.

resource "digitalocean_kubernetes_cluster" "doks" {

  name    = "do-challenge-k8s-cluster"
  region  = "lon1"
  version = "1.21.5-do.0"

  node_pool {
    name       = "autoscale-worker-pool"
    size       = "s-2vcpu-2gb"
    auto_scale = true
    min_nodes  = 1
    max_nodes  = 3
  }
}

Kubegres operator is installed in the cluster using the kubegres.yaml manifests.

data "kubectl_file_documents" "kubegres_operator_files" {
    content = file("${path.module}/manifests/kubegres.yaml")
}

resource "kubectl_manifest" "kubegres-operator" {
  for_each  = data.kubectl_file_documents.kubegres_operator_files.manifests
  yaml_body = each.value
}

K8s secret is created with pg_user and pg_pwd vars.

resource "kubernetes_secret" "pg-secret" {
  metadata {
    name = "pg-secret"
  }

  data = {
    "superUserPassword" = var.pg_user
    "replicationUserPassword" = var.pg_pwd
  }

  depends_on = [digitalocean_kubernetes_cluster.doks]
}

Postgres cluster is created with pg-cluster manifests.

resource "kubectl_manifest" "postgres-cluster" {
  yaml_body = "${file("${path.module}/manifests/pg-cluster.yaml")}"
  depends_on = [
    kubernetes_secret.pg-secret,
    kubectl_manifest.kubegres-operator,
    digitalocean_kubernetes_cluster.doks
  ]
}