fix(engine-b): address Codex round 5 (capacity release, activate-last, guest-ip poll, email unique)

- Failed run releases the instance (status=failed) so its quota stops counting
  against host capacity.
- CompleteProvisioning activates instance+order only AFTER onboarding tasks +
  credential delivery succeed.
- ConfigureNetwork polls until the guest has an address (no wrong Traefik route).
- customers.email unique + race-safe customer resolution in the intake action.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

app/Actions/StartCustomerProvisioning.php

@@ -25,13 +25,10 @@ class StartCustomerProvisioning
             return null; // already processed
         }
 
-        try {
-            [$order, $run] = DB::transaction(function () use ($event) {
-                $customer = Customer::query()->firstOrCreate(
-                    ['email' => $event['email']],
-                    ['name' => $event['name'] ?? $event['email'], 'stripe_customer_id' => $event['stripe_customer_id'] ?? null],
-                );
+        $customer = $this->resolveCustomer($event);
 
+        try {
+            [$order, $run] = DB::transaction(function () use ($event, $customer) {
                 $order = Order::create([
                     'customer_id' => $customer->id,
                     'plan' => $event['plan'],
@@ -61,4 +58,22 @@ class StartCustomerProvisioning
 
         return $order;
     }
+
+    /**
+     * Find or create the customer, race-safe against the unique email index so a
+     * concurrent first-time purchase can't create two customers for one email.
+     *
+     * @param  array{email:string,name:?string,stripe_customer_id:?string}  $event
+     */
+    private function resolveCustomer(array $event): Customer
+    {
+        try {
+            return Customer::query()->firstOrCreate(
+                ['email' => $event['email']],
+                ['name' => $event['name'] ?? $event['email'], 'stripe_customer_id' => $event['stripe_customer_id'] ?? null],
+            );
+        } catch (UniqueConstraintViolationException) {
+            return Customer::query()->where('email', $event['email'])->firstOrFail();
+        }
+    }
 }

app/Models/Order.php

@@ -40,9 +40,14 @@ class Order extends Model implements ProvisioningSubject
         return $this->morphMany(ProvisioningRun::class, 'subject');
     }
 
-    /** Runner hook: a failed provisioning run marks the order failed (no auto-refund). */
+    /**
+     * Runner hook: a failed provisioning run marks the order failed (no
+     * auto-refund) AND releases the reserved instance, so its quota stops
+     * counting against host capacity.
+     */
     public function onProvisioningFailed(): void
     {
         $this->update(['status' => 'failed']);
+        $this->instance()->update(['status' => 'failed']);
     }
 }

app/Provisioning/Steps/Customer/CompleteProvisioning.php

@@ -25,14 +25,12 @@ class CompleteProvisioning extends CustomerStep
         $order = $this->order($run);
         $instance = $this->instance($run);
 
-        $instance->update(['status' => 'active']);
-        $order->update(['status' => 'active']);
-
+        // Do all fallible work FIRST, so a failure can't leave the instance
+        // active. Onboarding tasks (idempotent) + credential delivery.
         foreach (['invite_users', 'upload_logo', 'connect_desktop', 'explore_apps'] as $key) {
             $instance->onboardingTasks()->firstOrCreate(['key' => $key], ['done' => false]);
         }
 
-        // Deliver the initial credentials, then scrub the transient password.
         $encrypted = $run->context('admin_password');
         if ($encrypted !== null) {
             Notification::route('mail', $order->customer->email)->notify(
@@ -41,6 +39,10 @@ class CompleteProvisioning extends CustomerStep
             $run->forgetContext('admin_password');
         }
 
+        // Activate LAST — only once everything above succeeded.
+        $instance->update(['status' => 'active']);
+        $order->update(['status' => 'active']);
+
         return StepResult::advance();
     }
 }

app/Provisioning/Steps/Customer/ConfigureNetwork.php

@@ -17,7 +17,8 @@ class ConfigureNetwork extends CustomerStep
 
     public function maxDuration(): int
     {
-        return 120;
+        // Above the 90s guest-ip own-deadline below.
+        return 150;
     }
 
     public function execute(ProvisioningRun $run): StepResult
@@ -29,12 +30,18 @@ class ConfigureNetwork extends CustomerStep
 
         $this->guest($pve, $run, 'hostnamectl set-hostname '.escapeshellarg($instance->subdomain));
 
-        // Capture the guest's own address so Traefik can route to the VM (not the host).
-        $ipOut = trim($pve->guestExec($node, $vmid, "hostname -I")['out-data'] ?? '');
+        // Capture the guest's own address so Traefik routes to the VM, not the
+        // host. Poll until it's available rather than advancing with no address.
+        $ipOut = trim($pve->guestExec($node, $vmid, 'hostname -I')['out-data'] ?? '');
         $guestIp = trim(strtok($ipOut, ' ') ?: '');
-        if ($guestIp !== '') {
-            $instance->update(['guest_ip' => $guestIp]);
+        if ($guestIp === '') {
+            if ($run->started_at !== null && $run->started_at->copy()->addSeconds(90)->isPast()) {
+                return StepResult::fail('guest_ip_unavailable');
+            }
+
+            return StepResult::poll(10, 'waiting for the guest to obtain an address');
         }
+        $instance->update(['guest_ip' => $guestIp]);
 
         // Allow HTTP/HTTPS, deny everything else (management stays on the tunnel).
         $pve->applyFirewall($node, $vmid, [

database/migrations/2026_07_25_080010_add_unique_email_to_customers_table.php

@@ -0,0 +1,24 @@
+<?php
+
+use Illuminate\Database\Migrations\Migration;
+use Illuminate\Database\Schema\Blueprint;
+use Illuminate\Support\Facades\Schema;
+
+return new class extends Migration
+{
+    public function up(): void
+    {
+        // One customer per email — the dashboard and broadcast auth resolve a
+        // customer by email, so duplicates would hide a customer's runs.
+        Schema::table('customers', function (Blueprint $table) {
+            $table->unique('email');
+        });
+    }
+
+    public function down(): void
+    {
+        Schema::table('customers', function (Blueprint $table) {
+            $table->dropUnique(['email']);
+        });
+    }
+};

tests/Feature/Provisioning/CustomerDataModelTest.php

@@ -18,12 +18,16 @@ it('resolves an Order as the polymorphic run subject', function () {
         ->and($order->getRouteKeyName())->toBe('uuid');
 });
 
-it('marks the order failed via the ProvisioningSubject hook', function () {
+it('marks the order and releases its instance via the ProvisioningSubject hook', function () {
     $order = Order::factory()->create(['status' => 'provisioning']);
+    $instance = Instance::factory()->create([
+        'order_id' => $order->id, 'customer_id' => $order->customer_id, 'status' => 'provisioning',
+    ]);
 
     $order->onProvisioningFailed();
 
-    expect($order->fresh()->status)->toBe('failed');
+    expect($order->fresh()->status)->toBe('failed')
+        ->and($instance->fresh()->status)->toBe('failed'); // quota no longer counts against capacity
 });
 
 it('enforces a unique stripe_event_id (idempotency)', function () {

tests/Feature/Provisioning/CustomerProvisioningEndToEndTest.php

@@ -14,7 +14,8 @@ it('provisions a paid order all the way to active (mocked)', function () {
     Notification::fake();
     Queue::fake(); // drive the runner manually, one step per advance
     $s = fakeServices();
-    $s['pve']->guestScript('status.php', 0, 'ok'); // deploy health + acceptance probe
+    $s['pve']->guestScript('status.php', 0, 'ok');
+    $s['pve']->guestScript('hostname -I', 0, '10.20.0.9'); // deploy health + acceptance probe
     // Fake defaults: tasks stopped/OK, guest agent up, cert reachable.
 
     Host::factory()->active()->create(['datacenter' => 'fsn', 'node' => 'pve', 'total_gb' => 1000]);
@@ -62,6 +63,7 @@ it('does not duplicate external resources when a step re-runs after a crash', fu
     Queue::fake();
     $s = fakeServices();
     $s['pve']->guestScript('status.php', 0, 'ok');
+    $s['pve']->guestScript('hostname -I', 0, '10.20.0.9');
 
     Host::factory()->active()->create(['datacenter' => 'fsn', 'node' => 'pve', 'total_gb' => 1000]);
     $order = Order::factory()->create(['status' => 'paid', 'plan' => 'start', 'datacenter' => 'fsn']);

tests/Feature/Provisioning/CustomerStepsTest.php

@@ -186,6 +186,12 @@ it('configures the guest network, captures the ip, and firewall', function () {
         ->and($s['pve']->firewallCalls)->toContain('101');
 });
 
+it('polls network config until the guest has an address', function () {
+    fakeServices(); // hostname -I returns empty by default → no address yet
+    ['run' => $run] = reservedRun();
+    expect(app(ConfigureNetwork::class)->execute($run)->type)->toBe('poll');
+});
+
 // 8. DeployApplicationStack
 it('deploys the stack, polling until healthy, without persisting the db password', function () {
     $s = fakeServices();