@kuckit/infra-gcp

GCP infrastructure provider for Kuckit applications. Deploys to Google Cloud Platform with:

- Cloud Run - Containerized application hosting
- Cloud SQL (PostgreSQL) - Managed database
- Memorystore (Redis) - Caching and sessions
- Artifact Registry - Container image storage
- VPC with Direct VPC Egress - Private networking

Prerequisites

- Pulumi CLI
- Google Cloud SDK (gcloud)
- GCP project with billing enabled
- Authenticated via gcloud auth login and gcloud auth application-default login

Installation

``bash
bun add -D @kuckit/infra-gcp


or


npm install -D @kuckit/infra-gcp
`

Usage

This package is used by the Kuckit CLI. After installation:

`bash


One-command deploy (init + deploy)

bunx kuckit infra up --env dev

Or step by step:

bunx kuckit infra init --provider gcp --project YOUR_GCP_PROJECT
bunx kuckit infra deploy --env dev

View status

bunx kuckit infra status --env prod

Destroy infrastructure

bunx kuckit infra destroy --env dev
`

How It Works

1. kuckit infra init - Creates base infrastructure without deploying the app:
- VPC with private subnets
- Cloud SQL PostgreSQL instance
- Memorystore Redis instance
- Artifact Registry repository
- Service accounts with minimal permissions

2. kuckit infra deploy - Builds and deploys your application:
- Builds Docker image using Cloud Build
- Pushes to Artifact Registry
- Deploys Cloud Run service with VPC connector
- Configures secrets and environment variables

Configuration

Infrastructure configuration is stored per-environment in .kuckit/:

| File | Environment | Description |
| ------------------------- | ----------- | ------------------------------- |
| .kuckit/infra.dev.json | Development | Dev environment settings |
| .kuckit/infra.prod.json | Production | Production environment settings |
| .kuckit/infra.json | Fallback | Legacy/default configuration |

Example configuration:

`json
{
"provider": "gcp",
"providerPackage": "@kuckit/infra-gcp",
"region": "us-central1",
"env": "dev",
"providerConfig": {
"gcpProject": "your-gcp-project"
}
}
`

$3

The infrastructure uses Pulumi project name kuckit-infra. Stack names follow the pattern {gcpProject}-{env}.

> ⚠️ Important: Changing the Pulumi project name causes resource recreation. See AGENTS.md for details.

Managing Configuration

`bash


Set custom domain

bunx kuckit infra config set appUrl https://app.example.com --env prod

Set Polar environment

bunx kuckit infra config set polarEnvironment production --env prod

List all config

bunx kuckit infra config list --env prod
`

Ejecting for Customization

If you need to customize the infrastructure beyond what the provider offers:

`bash
bunx kuckit infra eject
`

This copies the Pulumi code to your project's infra/ directory for full customization.

Environment Defaults

| Setting | Dev | Prod |
| ---------------- | ----------- | ----------- |
| Cloud Run CPU | 1 | 2 |
| Cloud Run Memory | 512Mi | 1Gi |
| Min Instances | 0 | 1 |
| Max Instances | 2 | 10 |
| Cloud SQL Tier | db-f1-micro | db-g1-small |

API Reference

This package exports:

`typescript
import { provider, gcpProvider } from '@kuckit/infra-gcp'
import type { GcpInfraConfig, GcpOutputs } from '@kuckit/infra-gcp'
`

The provider implements the KuckitInfraProvider interface from @kuckit/cli`.

Documentation

- Deployment Guide
- AGENTS.md - Provider internals and advanced configuration

License

MIT