@reyemtech/nimbus

Cloud-agnostic infrastructure abstractions for Pulumi. Write IaC once, deploy to AWS, Azure, or both. Cloud migration = change one config value.

Why

Every client environment follows the same pattern: network + cluster + DNS + secrets + platform. The only difference is the cloud provider. Nimbus extracts that pattern into reusable, type-safe factory functions.

• Cloud migration in one line — change cloud: "aws" to cloud: "azure"
• Multi-cloud active-active — deploy to AWS + Azure with a Global Load Balancer
• Dynamic provider loading — only the targeted SDK loads; missing providers give install instructions
• Cost-optimized defaults — fck-nat (~$3/mo vs $32/mo), spot instances, Auto Mode
• Escape hatches — every resource exposes nativeResource for cloud-specific access

Install

npm install @reyemtech/nimbus

Install provider SDKs for your target cloud(s):

npx @reyemtech/nimbus install aws          # @pulumi/aws
npx @reyemtech/nimbus install azure        # @pulumi/azure-native
npx @reyemtech/nimbus install aws azure    # both

Quick Start

Scaffold a new project (handles pulumi new, npm install, and provider setup):

npx @reyemtech/nimbus new my-infra aws
cd my-infra
pulumi up

Or start from scratch:

import { createNetwork, createCluster, createPlatformStack } from "@reyemtech/nimbus";
import type { INetwork, ICluster } from "@reyemtech/nimbus";

const network = await createNetwork("prod", {
  cloud: "aws",
  cidr: "10.0.0.0/16",
  natStrategy: "fck-nat",
}) as INetwork;

const cluster = await createCluster("prod", {
  cloud: "aws",
  nodePools: [
    { name: "system", instanceType: "t4g.small", minNodes: 2, maxNodes: 3 },
    { name: "workers", instanceType: "c6a.large", minNodes: 1, maxNodes: 10, spot: true },
  ],
  providerOptions: { aws: { autoMode: true } },
}, network) as ICluster;

createPlatformStack("prod", { cluster, domain: "example.com" });

See docs/examples.md for Azure, multi-cloud, and standalone resource patterns.

CLI

nimbus new <name> <template>     # Scaffold a project from a template
nimbus install <provider> [...]  # Install cloud provider SDKs
nimbus check                     # Show which providers are installed

Templates

Template	Description
empty	Blank scaffold with TODO placeholders
minimal-aws	State backend + Secrets Manager (no cluster)
minimal-azure	State backend + Key Vault (no cluster)
aws	Full stack: VPC + EKS + Route 53 + Secrets + Platform
azure	Full stack: VNet + AKS + Azure DNS + Key Vault + Platform
multi-cloud	AWS + Azure active-active with Global Load Balancer

Day-to-Day Operations

After scaffolding, use standard Pulumi commands to manage your infrastructure:

pulumi preview    # Preview changes before deploying
pulumi up         # Deploy infrastructure
pulumi refresh    # Sync state with actual cloud resources
pulumi destroy    # Tear down all resources
pulumi config set key value   # Set stack configuration
pulumi stack output           # View stack outputs

Architecture

@reyemtech/nimbus
├── factories/      # Cloud-agnostic factory functions (primary API)
├── types/          # CloudProvider, CloudTarget, tags, errors, validation
├── network/        # VPC (AWS), VNet (Azure) + NAT + CIDR utilities
├── cluster/        # EKS, AKS (+ Auto Mode, virtual nodes, spot)
├── dns/            # Route 53, Azure DNS
├── secrets/        # AWS Secrets Manager, Azure Key Vault
├── state/          # Pulumi state backend (S3, Azure Blob) with BCDR
├── platform/       # Helm: Traefik, cert-manager, External DNS, ArgoCD, Vault, ESO
├── global-lb/      # DNS-based multi-cloud routing (active-active, failover, geo)
├── cli.ts          # CLI (nimbus new/install/check)
├── cli/            # CLI templates for project scaffolding
├── database/       # (interfaces only — RDS, Azure DB, PXC, CNPG)
├── cache/          # (interfaces only — ElastiCache, Azure Cache, Redis Helm)
├── storage/        # (interfaces only — S3, Blob, GCS)
└── queue/          # (interfaces only — SQS, Service Bus, NATS)

Module Status

Module	AWS	Azure	GCP
Network	VPC + fck-nat/managed NAT	VNet + NAT Gateway	Planned
Cluster	EKS + Auto Mode	AKS + virtual nodes	Planned
DNS	Route 53	Azure DNS	Planned
Secrets	Secrets Manager	Key Vault	Planned
State	S3 + DynamoDB locking + replication	Blob + GRS	Planned
Platform	Helm (provider-agnostic)	Helm (provider-agnostic)	Helm (provider-agnostic)
Global LB	Route 53 health-checked routing	Planned	—

Documentation

• API Reference — Factory functions, provider options, direct cloud functions
• Examples — Single-cloud, multi-cloud, standalone resources, BCDR
• Utilities — CIDR helpers, validation, tags, error handling

Development

npm install
npm run typecheck      # Type checking
npm run lint           # ESLint
npm run format         # Prettier
npm test               # Vitest (unit tests)
npm run test:coverage  # With 80% coverage thresholds
npm run build          # ESM + CJS dual build
npm run docs:api       # TypeDoc API docs

License

MIT