Local development environments#

Goal. New engineers clone the repo, run one command, and have a working local database with enough data to develop against. No “ask Slack for a SQL dump.”

Why Orchard. A single authoritative scenario replaces the tribal-knowledge approach. The scenario lives in the repo and evolves with the schema.

Pattern: one “dev” scenario per service#

scenarios/
└── dev.hcl

dev.hcl is the canonical “everything you need to boot up locally.” Engineers run it after applying migrations.

Example#

scenario "dev" {
  required_providers {
    postgres = { source = "builtin/postgres" }
    exec     = { source = "builtin/exec" }
  }

  variable "dsn" {
    default = "postgres://localhost:5432/myapp_dev?sslmode=disable"
  }

  provider "postgres" {
    dsn = var.dsn
  }

  provider "exec" {}

  # Confirm the database is reachable before we try to seed it.
  action "exec_run" "check_db" {
    command = "psql"
    args    = [var.dsn, "-c", "SELECT 1"]
  }

  component "admin" "seed_admin" {
    source = "../components/user.hcl"
    inputs = {
      email = "admin@local.dev"
      role  = "admin"
    }
  }

  component "fixtures" "dev_data" {
    source = "../components/dev_fixtures.hcl"
    inputs = {
      admin_id = component.admin.seed_admin.id
    }
  }

  output "admin_email"   { value = component.admin.seed_admin.email }
  output "login_url"     { value = "http://localhost:3000/login" }
}

The engineer runs:

make db-setup              # runs migrations
orchard run scenarios/dev.hcl

And sees:

admin_email = admin@local.dev
login_url   = http://localhost:3000/login

Pattern: idempotent vs destructive#

Local dev scenarios are typically destructive — they assume a fresh database. That’s fine; document it. The alternative is writing idempotent SQL (INSERT ... ON CONFLICT DO UPDATE), which works but complicates the scenario.

For most teams: “blow it away and re-seed” is the right default. Add a make db-reset that drops and recreates the database, then runs the dev scenario.

Tips#

Keep it fast. Local dev is a daily ritual — the scenario should finish in seconds, not minutes. Pare down the data to what’s needed for day-to-day work.
Split heavy data into opt-in scenarios. Keep dev.hcl minimal and lean; offer dev_heavy.hcl for engineers working on reporting or perf.
Document required env vars in one place. If the scenario uses --var stripe_api_key=..., say so in the repo’s README — Orchard will tell you about the missing variable, but a pointer up front is kinder.