Golde

GitOps based infrastructure as code and deployment tool.

What is it?

• Simple declarative infrastructure as code tool focused on common use cases. Starting with self-hosted infrastructure and major cloud providers.

• Git-based application deployments. All resources have specific branch ownership. Covers common use cases like containerized applications and static websites.

What is it not?

• Terraform. It lacks the fancy configuration language, third-party providers, package manager, plugins and other features.

• Kubernetes. It is not a platform to abstract away all infrastructure details. Helm charts, complex networking are not in scope.

• Fly/Vercel. It will not provide fully managed services like databases, storage, CI/CD, etc.

Why it is better

Than terraform

• Easier to learn, no custom configuration language, no package manager, plugin system and different extensions tfstate, tfvars, etc.

• Easy to create higher-level abstractions using standard languages like TypeScript, JavaScript, Python, etc. No need for another orchestration layer like terragrunt

• Map directly to common Git based deployment workflows, eliminate need for multiple workspaces or env specific directories

• Deep schema validation and type safety for config. Planning performs permission simulations and existence checks. Reduces chances of a failed apply.

• Config has a better self-documenting structure, name of resource is part of config, not a separate property. Config maps directly to physical resources unlike terraform where resources are often meta-resources created for the sake of partitioning config.

• Properly open-sourced, so you can contribute and extend it without worrying about rug pulls. Monolithic structure and TypeScript make it very easy to contribute.

Planned premium features

• State locking and concurrency control for state
• Cross project state references
• Managed docker registry
• Infrastructure agent to manage dedicated infrastructure
• Uptime monitoring and alerting
• Resource previews and infrastructure diagrams
• Feature flags and config management
• Logging and monitoring on top of deployed infrastructure

Ok, give me examples

Serverless next.js deploy

• Create necessary infrastructure in AWS
• Package artifacts and upload to s3
• Deploy next.js app on master branch and branches starting feature/*
• Pruning of feature branches resources

import type {Config} from '@golde/cli';

const config: Config = {
  name: "example-aws-lambda-next-js",
  tags: {
    Project: "GoldeExamples",
    Example: "example-aws-lambda-next-js",
    Branch: "{{ git.BRANCH_NAME }}",
  },
  providers: {
    golde: {
      apiKey: "{{ env.GOLDE_API_KEY }}",
    },
    aws: {
      accessKeyId: "{{ env.AWS_ACCESS_KEY_ID }}",
      secretAccessKey: "{{ env.AWS_SECRET_ACCESS_KEY }}",
    },
  },
  resources: {
    aws: {
      s3: {
        bucket: {
          "example-aws-lambda-next-js": {
            branch: "master",
          },
          "example-aws-lambda-next-js-{{ git.BRANCH_SLUG }}": {
            branchPattern: "feature/*",
          }
        },
        object: {
          "lambda-{{ git.BRANCH_HASH }}.zip": {
            includes: [
              {from: ".next/standalone", to: "."},
              {from: "bin/run.sh", to: "."},
              {from: "public", to: "public"},
            ],
            branch: "master",
            bucketName: "{{ state.aws.s3.bucket.example-aws-lambda-next-js.name }}",
          },
          "lambda-{{ git.BRANCH_SLUG }}-{{ git.BRANCH_HASH }}.zip": {
            includes: [
              {from: ".next/standalone", to: "."},
              {from: "bin/run.sh", to: "."},
              {from: "public", to: "public"},
            ],
            branchPattern: "feature/*",
            bucketName: "{{ state.aws.s3.bucket.example-aws-lambda-next-js-{{ git.BRANCH_SLUG }}.name }}",
          },
        },
      },
      iam: {
        role: {
          "example-aws-lambda-next-js": {
            assumeRolePolicy: {
              "Version": "2012-10-17",
              "Statement": [
                {
                  "Action": "sts:AssumeRole",
                  "Principal": {
                    "Service": "lambda.amazonaws.com"
                  },
                  "Effect": "Allow",
                  "Sid": ""
                },
              ]
            },
            inlinePolicy: {
              "Version": "2012-10-17",
              "Statement": [
                {
                  "Action": [
                    "logs:PutLogEvents",
                  ],
                  "Resource": "arn:aws:logs:*:*:*",
                  "Effect": "Allow"
                }
              ]
            }
          }
        },
      },
      cloudwatch: {
        logGroup: {
          "/aws/lambda/example-aws-lambda-next-js": {
            branch: "master",
            retentionInDays: 60,
          },
          "/aws/lambda/example-aws-lambda-next-js-{{ git.BRANCH_SLUG }}": {
            branchPattern: "feature/*",
            retentionInDays: 14,
          },
        },
      },
      lambda: {
        function: {
          "example-aws-lambda-next-js": {
            branch: "master",
            packageType: "Zip",
            description: "Example AWS Lambda Next.js",
            runtime: "nodejs20.x",
            handler: "index.handler",
            memorySize: 512,
            timeout: 30,
            roleArn: "{{ state.aws.iam.role.example-aws-lambda-next-js.arn }}",
            code: {
              s3Bucket: "{{ state.aws.s3.bucket.example-aws-lambda-next-js.name }}",
              s3Key: "lambda-{{ git.BRANCH_HASH }}.zip",
            },
            loggingConfig: {
              logGroupName: "{{ state.aws.cloudwatchLogGroup./aws/lambda/example-aws-lambda-next-js.name }}",
            }
          },
          "example-aws-lambda-next-js-{{ git.BRANCH_SLUG }}": {
            branchPattern: "feature/*",
            packageType: "Zip",
            description: "Feature branch example AWS Lambda Next.js",
            runtime: "nodejs20.x",
            handler: "index.handler",
            memorySize: 256,
            timeout: 30,
            roleArn: "{{ state.aws.iam.role.example-aws-lambda-next-js.arn }}",
            code: {
              s3Bucket: "{{ state.aws.s3.bucket.example-aws-lambda-next-js-{{ git.BRANCH_SLUG }}.name }}",
              s3Key: "lambda-{{ git.BRANCH_SLUG }}-{{ git.BRANCH_HASH }}.zip",
            },
            loggingConfig: {
              logGroupName: "{{ state.aws.cloudwatchLogGroup./aws/lambda/example-aws-lambda-next-js-{{ git.BRANCH_SLUG }}.name }}",
            }
          },
        }
      }
    }
  },
};

export default config

But Marcin, how is this any different than CloudFormation or Terraform? Because config is not YAML or JSON, it is actually very easy to build abstractions.

import type {Config} from '@golde/cli';
import {createAWSDeployment} from '@golde/nextjs';

const name = "example-aws-lambda-next-js"
const region = "us-east-1"
const aws = createAWSDeployment({
  name,
  region,
  customsAWS: {
    // Add any custom aws beside standard next js deployment
  }
});

const config: Config = {
  name: "example-aws-lambda-next-js",
  tags: {
    Project: "GoldeExamples",
    Example: "example-aws-lambda-next-js",
    Branch: "{{ git.BRANCH_NAME }}",
  },
  providers: {
    golde: {
      apiKey: "{{ env.GOLDE_API_KEY }}",
    },
    aws: {
      accessKeyId: "{{ env.AWS_ACCESS_KEY_ID }}",
      secretAccessKey: "{{ env.AWS_SECRET_ACCESS_KEY }}",
    },
  },
  resources: {
    aws,
  },
};

export default config