Provisioning a New Instance

A fresh FlexPoint Portal instance (Forge staging / candidate / production) is bootstrapped with a single command:

php artisan flexpoint:provision --with-admin --force

This seeds the foundational relational data and dispatches the MeridianLink (MLM) identity imports onto their priority queues.

What it does, in order

App\Console\Commands\ProvisionInstanceCommand

1. Permissions + system roles (+ role→permission pivots) — seeded synchronously via IdentitySeeder.
2. Notification templates + workflows — seeded synchronously via NotificationSeeder.
3. Bootstrap Super Admin (ADMIN_EMAIL / test@example.com) — seeded synchronously via AdminUserSeeder, only when --with-admin is passed.
4. Originating-company sync → dispatched to queue mlm-companies.
5. Internal-user sync → dispatched to queue identity.

Roles must be seeded synchronously first: both MLM user-sync jobs assign portal roles by key lookup (Role::where('key', …)->value('id')), which silently no-ops if the role rows do not yet exist. Seeding inline before dispatch removes that race. The originating-company job then fans out staggered ImportExternalUsers jobs (queue mlm-users) for every company it discovers.

Flags

All seeders are idempotent and faker-free, so the command is safe to re-run on every deploy.

Queues

A Horizon daemon drains these queues in priority order.

Forge setup

Deploy hook (Site → Deployments):

cd $FORGE_SITE_PATH/apps/flex-hub-api
composer install --no-dev --no-interaction --prefer-dist --optimize-autoloader
php artisan migrate --force
php artisan config:cache && php artisan route:cache && php artisan event:cache
php artisan horizon:terminate

Forge daemon (Server → Daemons) — drains all priority queues:

php artisan horizon

One-time, on a brand-new instance (after the first deploy):

php artisan flexpoint:provision --with-admin

Fallback worker if Redis/Horizon is unavailable (database driver):

php artisan queue:work --queue=identity,mlm-companies,mlm-users,default