Cloud Seed

Using on a project:

Add the package as a dev dependency (note you'll need to authenticate with github packages).

npm install -D @space48/cloud-seed

Define your first cloud function:

src/myFirstFunction.ts

import { HttpFunction } from "@google-cloud/functions-framework";
import { GcpConfig } from "@space48/cloud-seed";

const myFunction: HttpFunction = (req, res) => {
  console.log("Hello World");
  return res.sendStatus(200);
};

export default myFunction;

export const runtimeConfig: GcpConfig = {
  cloud: "gcp",
  type: "http",
  public: true,
};

Then run your build command:

npx @space48/cloud-seed build

To apply the terraform config you'll be able to run the following:

terraform chdir=[buildDir] init
terraform chdir=[buildDir] plan -out=plan
terraform chdir=[buildDir] apply plan

Running functions locally for testing

You can start the local development server for a function by running:

npx @space48/cloud-seed run src/myFirstFunction.ts --env environment

The development server will build the function and expose it on http://localhost:3000/ so you can trigger it with a HTTP request. Any changes you make to the source code of the function will rebuild it and restart the server, allowing you to keep testing without having to manually run any additional commands.

Tips and tricks:

• The development server does not require access to a cloud project to run the function locally. However, if your function code accesses any cloud resources, such as message queues or databases, you will need to authenticate with the cloud to allow it. For example, you can authenticate with GCP by running the gcloud auth login command in gcloud CLI.
• The development server runs a single cloud function in isolation. If you want to test multiple functions, you can run multiple development servers on different ports using the --port option. Keep in mind that this won't automatically connect the functions and allow then to interract with each other. Your function code will have to be aware that it is running in development and adjust the endpoints it uses accordingly. You can use the CLOUD_SEED_ENVIRONMENT environment variable to write conditional code for running in the development environment.
• Any cloud function can be triggered by a HTTP request, regardless of its type. However, some function types may require the request data to follow a specific format or use a specific encoding. Refer to the official documentation for the cloud service you're using for details on how to format your function input.
• The development server automatically loads the environment variables you have defined in cloudseed.json for the provided environment and exposes them to the function. You can use this to configure a dedicated development environment, with appropriate configuration, to be shared by anyone running the project locally. Additionally, any environment variables with the same names defined in your environment will override the values from the configuration file. This is useful to temporarily modify the configuration for testing without having to modify the cloudseed.json file.

How does this work?

Each Cloud Function you wish to define can live anywhere within the src directory, how you structure this is up to you.

The only thing you need to define for your function to work is the named export: runtimeConfig. Take a look at the RuntimeConfig type to see what options you have (thats in runtime.ts).

Supported function types:

HTTP Functions:

import { HttpFunction } from "@google-cloud/functions-framework";
import type { GcpConfig } from "@space48/cloud-seed";

const myFunction: HttpFunction = (req, res) => {
  console.log("Hello World");
  return res.sendStatus(200);
};

export default myFunction;

export const runtimeConfig: GcpConfig = {
  cloud: "gcp",
  type: "http",
  // Configure if the cloud function should be publically executable or not.
  public: true,
};

PubSub topic subscriptions:

import { CloudEventFunction } from "@google-cloud/functions-framework";
import type { GcpConfig } from "@space48/cloud-seed";

const fn: CloudEventFunction (data) => {
  console.log("This is a event triggered function", data);
};

export default fn;

export const runtimeConfig: GcpConfig = {
  cloud: "gcp",
  type: "event",
  // By defining the topicName it will create the topic for you.
  topicName: "hello-world",
};

Scheduled functions using pub/sub topic:

import { CloudEventFunction } from "@google-cloud/functions-framework";
import type { GcpConfig } from "@space48/cloud-seed";

const fn: CloudEventFunction = (data) => {
  console.log("This is a scheduled triggered function", data);
};

export default fn;

export const runtimeConfig: GcpConfig = {
  cloud: "gcp",
  type: "schedule",
  schedule: "* * * * *",
};

Scheduled functions without using pub/sub topic:

Supported only in gen 2 supported versions of cloud-seed

import { CloudEventFunction } from "@google-cloud/functions-framework";
import type { GcpConfig } from "@space48/cloud-seed";

const fn: CloudEventFunction = (data) => {
  console.log("This is a scheduled triggered function, triggered directly from the scheduled job", data);
};

export default fn;

export const runtimeConfig: GcpConfig = {
  cloud: "gcp",
  schedule: "* * * * *",
  type: "scheduledJob",
  version: "gen2",
  attemptDeadline: "1800s", // optional, this is set as 3 minutes by default
};

Queues:

import { HttpFunction } from "@google-cloud/functions-framework";
import type { GcpConfig } from "@space48/cloud-seed";

const fn: HttpFunction = (req, res) => {
  console.log(req.body);
  return res.sendStatus(200);
};

export default fn;

export const runtimeConfig: GcpConfig = {
  cloud: "gcp",
  type: "queue",
};

Firestore document triggers:

import { CloudEventFunction } from "@google-cloud/functions-framework";
import type { GcpConfig } from "@space48/cloud-seed";

const fn: CloudEventFunction = (data) => {
  console.log("This is a firestore triggered function", data);
};

export default fn;

export const runtimeConfig: GcpConfig = {
  cloud: "gcp",
  type: "firestore",
  document: "collection/{doc}",
  // Optional event type (defaults to 'write').
  firestoreEvent: "create"
};

Cloud Storage triggers

import { CloudEventFunction } from "@google-cloud/functions-framework";
import type { GcpConfig } from "@space48/cloud-seed";

const fn: CloudEventFunction = (data) => {
  console.log("This is a cloud storage triggered function", data);
};

export default fn;

export const runtimeConfig: GcpConfig = {
  cloud: "gcp",
  type: "storage",
  bucket: {
    default: "myBucket",
    // optional environment-specific buckets
    environmentSpecific: {
      production: "myProdBucket",
      ...
    },
  },
  // Optional event type (defaults to 'finalize').
  storageEvent: "finalize"
};

Environment-level function configuration

You can override any value in runtimeConfig for a specific environment by including an environmentOverrides object. For example:

import { GcpConfig } from "@space48/cloud-seed";

export const runtimeConfig: GcpConfig = {
  cloud: "gcp",
  type: "http",
  public: true,
  minInstances: 0,
  environmentOverrides: {
    production: {
      minInstances: 1,
    },
  },
};

Setting up a config file:

You can set the cloud seed config by adding a cloudseed.json file in the project root directory. An example is provided below:

{
  "$schema": "./node_modules/@space48/cloud-seed/schemas/cloudseed.schema.json",
  "default": {
    "cloud": {
      "gcp": {
        "region": "europe-west2"
      }
    },
    "buildConfig": {
      "dir": "./src",
      "outDir": "./.build",
    },
    "secretVariableNames": [
      "apiKey1",
      "apiKey2"
    ]
  },
  "environmentOverrides": {
    "staging": {
      "cloud": {
        "gcp": {
          "project": "example-project-staging"
        }
      },
      "tfConfig": {
        "backend": {
          "type": "gcs",
          "backendOptions": {
            "bucket": "example-backend-bucket",
            "prefix": "path/to/staging/statefile/directory"
          }
        }
      },
      "runtimeEnvironmentVariables": {
        "FOO": "Staging1",
        "BAR": "Staging2"
      }
    },
    "production": {
      "cloud": {
        "gcp": {
          "project": "example-project-production"
        }
      },
      "tfConfig": {
        "backend": {
          "type": "gcs",
          "backendOptions": {
            "bucket": "example-backend-bucket",
            "prefix": "path/to/production/statefile/directory"
          }
        }
      },
      "runtimeEnvironmentVariables": {
        "FOO": "Prod1",
        "BAR": "Prod2"
      }
    }
  }
}

JavaScript API

This package can also be called via a JS API.

import { build, BaseConfig } from "@space48/cloud-seed";
import GoogleProvider from "@cdktf/provider-google";
import { GcsBackend, TerraformStack } from "cdktf";
import { Construct } from "constructs";

class CustomStack extends TerraformStack {
  constructor(scope: Construct, id: string, options: BaseConfig) {
    super(scope, id);
    options.tfConfig.backend.type === "gcs" && new GcsBackend(this, {
      bucket: options.tfConfig.backend.backendOptions.bucket,
      prefix: options.tfConfig.backend.backendOptions.prefix + "-custom-stack",
    });
    new GoogleProvider.GoogleProvider(this, "Google", {
      region: options.cloud.gcp.region,
      project: options.cloud.gcp.project,
    });

    /**
     * Example infrastructure
     */
    new GoogleProvider.StorageBucket(this, "CustomBucket", {
      name: `my-custom-bucket-${process.env.ENVIRONMENT}`,
      location: options.cloud.gcp.region.toUpperCase(),
      storageClass: "STANDARD",
      uniformBucketLevelAccess: true,
    });
  }
}

// build() is the same as per the CLI command, and returns the parsed Cloud Seed config and the app construct
const { config, app } = build({ environment: process.env.ENVIRONMENT });

new CustomStack(app, "CustomStack", config);

app.synth();

Cloud functions gen 2

It is recommended by google to use the gen 2 functions wherever possible.

The cloud-seed v1.3.6-alpha is an upgraded version to use gen 2 functions, which is still in alpha stage. This is currently being used in Sneakers-n-stuff.

If you upgrade a project, that is using an older version of cloud-seed (ex:v1.3.0 is the latest stable version) to the 1.3.6-alpha version, all the existing cloud functions (which use older csktf npm modules) will be deleted and recreated. This can lead to an issue as functions/topics/secrets can't be created with the same name for 7 days, when they get deleted. Therefore upgrading a project that uses cloud-seed, has to be done cautiously.

But for a new project it is recommended to use the 1.3.6-alpha version with gen 2 functions support.

Add a cloud 2 function

The structure is similar to what is explained above. But in runtime config add version: "gen2"

export const runtimeConfig: GcpConfig = {
  runtime: "nodejs18",
  cloud: "gcp",
  type: "schedule",
  schedule: "0 0 * * *",
  memory: 2048,
  timeout: 540,
  version: "gen2",
};

If a version is not given, the function will be automatically added as a gen 1.

prerequisites

• python and make should be installed in the build server

1. add to the bitbucket pipeline build step

 - step: &build
        name: "Build"
        caches:
          - npm
        script:
          # Install Python and Make
          - chmod +x ./bitbucket-pipelines/install_python_and_make.sh
          - ./bitbucket-pipelines/install_python_and_make.sh

2. Add a new bash script: install_python_and_make.sh

#!/bin/bash
apt update && apt -y install python3 && apt -y install build-essential

• Enable Eventarc API in GCP as gen2 cloud functions are built on Cloud Run and Eventarc

Licence and acknowledgements

This project is licenced under the MIT License.

The authors wish to acknowledge our collaboration with the open-source Cloud Seed project by user MNahad, and that certain features in our project are derived from it.