pulumi-defang-gcp 2.0.0a23

Deploy containerized services to GCP with Pulumi.

Defang Pulumi Provider

The Pulumi Provider for Defang — Take your app from Docker Compose to a secure and scalable cloud deployment with Pulumi.

Example usage

The examples below use AWS. GCP and Azure follow the same pattern — just swap the package name (e.g. defang-aws → defang-gcp). Complete working samples for all clouds and languages are in the ./examples directory.

{{< chooser language "typescript,python,go,dotnet,yaml" >}} {{% choosable language typescript %}}

import * as pulumi from "@pulumi/pulumi";
import * as defang_aws from "@defang-io/pulumi-defang-aws";

const awsDemo = new defang_aws.Project("aws-demo", {services: {
    app: {
        image: "nginx",
        ports: [{
            target: 80,
            mode: "ingress",
            appProtocol: "http",
        }],
    },
}});
export const endpoints = awsDemo.endpoints;

{{% /choosable %}}

{{% choosable language python %}}

import pulumi
import pulumi_defang_aws as defang_aws

aws_demo = defang_aws.Project("aws-demo", services={
    "app": {
        "image": "nginx",
        "ports": [{
            "target": 80,
            "mode": "ingress",
            "app_protocol": "http",
        }],
    },
})
pulumi.export("endpoints", aws_demo.endpoints)

{{% /choosable %}}

{{% choosable language go %}}

package main

import (
	defangaws "github.com/DefangLabs/pulumi-defang/sdk/v2/go/defang-aws"
	"github.com/DefangLabs/pulumi-defang/sdk/v2/go/defang-aws/compose"
	"github.com/pulumi/pulumi/sdk/v3/go/pulumi"
)

func main() {
	pulumi.Run(func(ctx *pulumi.Context) error {
		awsDemo, err := defangaws.NewProject(ctx, "aws-demo", &defangaws.ProjectArgs{
			Services: compose.ServiceConfigMap{
				"app": &compose.ServiceConfigArgs{
					Image: pulumi.String("nginx"),
					Ports: compose.ServicePortConfigArray{
						&compose.ServicePortConfigArgs{
							Target:      pulumi.Int(80),
							Mode:        pulumi.String("ingress"),
							AppProtocol: pulumi.String("http"),
						},
					},
				},
			},
		})
		if err != nil {
			return err
		}
		ctx.Export("endpoints", awsDemo.Endpoints)
		return nil
	})
}

{{% /choosable %}}

{{% choosable language dotnet %}}

using System.Collections.Generic;
using System.Linq;
using Pulumi;
using DefangAws = DefangLabs.DefangAws;

return await Deployment.RunAsync(() =>
{
    var awsDemo = new DefangAws.Project("aws-demo", new()
    {
        Services =
        {
            { "app", new DefangAws.Compose.Inputs.ServiceConfigArgs
            {
                Image = "nginx",
                Ports = new[]
                {
                    new DefangAws.Compose.Inputs.ServicePortConfigArgs
                    {
                        Target = 80,
                        Mode = "ingress",
                        AppProtocol = "http",
                    },
                },
            } },
        },
    });

    return new Dictionary<string, object?>
    {
        ["endpoints"] = awsDemo.Endpoints,
    };
});

{{% /choosable %}}

{{% choosable language yaml %}}

name: defang-aws
runtime: yaml
description: Example using defang-aws to deploy services to AWS

plugins:
  providers:
    - name: defang-aws
      path: ../../bin

resources:
  aws-demo:
    type: defang-aws:index:Project
    properties:
      services:
        app:
          image: nginx
          ports:
            - target: 80
              mode: ingress
              appProtocol: http

outputs:
  endpoints: ${aws-demo.endpoints}

{{% /choosable %}} {{< /chooser >}}

Components

Each provider (defang-aws, defang-gcp, defang-azure) exposes the same component palette:

• Project — the recommended entry point. Takes a full services map (Compose-style) and provisions shared infrastructure (VPC, networking, DNS, load balancers, build pipelines) alongside each service.
• Service — a single container service. Standalone use is image-only: image must refer to a pre-built image. Build-from-source is a Project responsibility because it needs the shared build pipeline (Artifact Registry + Cloud Build on GCP, ECR + CodeBuild on AWS, ACR on Azure).
• Postgres / Redis — managed database / cache components. Can be used standalone or as part of a Project.
• Build (AWS only) — an image-build resource used internally by Project.

Managed components (Service, Postgres, Redis) share one implementation between standalone and project-scoped use. Standalone instantiations skip the shared-infra provisioning and therefore don't support features that depend on it (VPC access, build-from-source, external load-balancer wiring).

Resource naming

When creating Pulumi resources, these Pulumi components will only specify the logical name, which is either the service name (e.g. app) or a name to describe the resource's role (e.g. shared-vpc, ecr-public). The physical name of the underlying cloud resource is determined by the Pulumi engine. By default, this will be the logical name followed by a hyphen and 7 random hex characters (e.g. app-abc1234). To control the physical name, configure autonaming rules in the Pulumi configuration files, either globally or per resource type. See the Pulumi autonaming docs for details. See the CD code for examples of how autonaming is used in Defang.

Installation and Configuration

See our Installation and Configuration docs

Development

See the Contributing doc.