Recipes — AWS (the whole cloud, one block)

One aws block runs the whole local cloud as a single process: S3, SQS, SNS, DynamoDB (with Streams), Lambda, EventBridge, KMS, SSM Parameter Store, and Secrets Manager — pure Go, one endpoint, with a web console at /_console. Buckets, queues, tables, and functions are declared inside the block and converged on boot.

• The block
• Point your SDK at it
• Queues, topics, fanout
• DynamoDB
• Lambda + EventBridge
• Config & secrets
• The console

The block

module aws "local" {
  port = 4566

  bucket "uploads" { versioning = true }

  queue "emails" {
    dlq = "auto"
    max_receives = 5
  }
  queue "orders.fifo" { # FIFO names end in .fifo
    fifo = true
    content_dedup = true
  }

  topic "signups" {
    subscribe {
      queue = "emails"
      raw = true
    }
  }

  table "sessions" {
    key = "session_id:S"
    ttl = "expires_at"
  }
}

Everything inside is created (or updated) when the instance boots, and it all persists to disk across reaps like any other engine. There’s no version = — the services track current AWS APIs.

Point your SDK at it

Reference module.local.url from an app block and every AWS SDK picks it up through the standard endpoint variables:

app "api" {
  run = "go run ."
  env = {
    AWS_ENDPOINT_URL_S3  = module.local.url    # boots + holds the aws instance
    BUCKET               = "uploads"        # resources are plain names
  }
}

export AWS_ENDPOINT_URL=$(doze env | grep -o 'http://aws[^"]*' | head -1)
export AWS_ACCESS_KEY_ID=test AWS_SECRET_ACCESS_KEY=test AWS_REGION=us-east-1
aws s3 ls
aws sqs get-queue-url --queue-name emails

One endpoint serves every service — the gateway routes by protocol, exactly like the standalone doze-aws binary.

Queues, topics, fanout

queue "emails" {
  visibility_timeout = "30s"
  retention          = "96h"     # Go durations; days aren't a unit, use hours
  dlq                = "auto"    # creates emails-dlq + redrive policy
  max_receives       = 5
}

topic "signups" {
  subscribe { # SNS→SQS fanout
    queue = "emails"
    raw = true
  }
}

DynamoDB

table "sessions" {
  key = "session_id:S"       # "name:type" — add a range key with sort = "at:N"
  ttl = "expires_at"
}

Streams are on — a Lambda can subscribe to table changes.

Lambda + EventBridge

function "resize" {
  code    = "./functions/resize"    # a dir with a provided.al2 bootstrap binary
  timeout = 10
}

rule "order_placed" {
  pattern = "{\"source\":[\"shop.checkout\"]}"   # a literal JSON string
  targets = ["lambda:resize", "queue:emails"]    # kind:name shorthand
}

Functions scale to zero like everything else — cold on first invoke, warm for a while, then asleep. PutEvents on the default bus fans out to every matching rule target.

Config & secrets

key       "app_key"     {}                          # KMS
parameter "feature_flag" { value = "on" }           # SSM Parameter Store
secret    "db_password"  { value = "local-only" }   # Secrets Manager

The console

Open the instance in the dash (o) or browse to http://<endpoint>/_console: every service has a page — browse buckets and objects, peek and purge queues, scan tables, invoke functions, decrypt nothing (KMS values stay masked) — plus a live Traffic tape of every API call your app makes, each replayable as a curl.

Dev-grade by design: fast and faithful enough to build and test against, not a replacement for real AWS in production.