PostgreSQL rejects encrypted (base64) strings in jsonb columns. The
encrypted:array cast produces non-JSON output that must be stored as text.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

- WhatsAppService now reads credentials from messaging_channels table
  (encrypted:array cast) instead of config/env vars
- New MessagingSettings Livewire page to configure credentials from UI
- Webhook verify token also read from DB
- API Keys page auto-pushes MESSAGING_HUB_URL + KEY to instances
  via CapRover env vars on generate/regenerate
- Sidebar shows messaging sub-nav when on messaging pages
- Removed hardcoded whatsapp config from manager.php
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Full messaging hub that instances call via API to send WhatsApp messages:
- Meta Graph API v25.0 integration (text, template, image, document)
- API endpoint POST /api/messaging/send (auth via X-Messaging-Key header)
- Per-instance API keys with rate limiting
- WhatsApp webhook for delivery status updates
- Admin UI: dashboard stats, compose, message log, API key management
- Sidebar nav integration
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

- Switch logs from broken POST /appDefinitions/logs to working GET /appData/{app}
- Add x-namespace header to auth call
- Add activate() method to transition trial/suspended → active
- Add Activate button in instance-show view
- Rename "Container Logs" to "Build Logs" (reflects actual content)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

- Client portal: public page for clients (auth by instance UUID)
  Shows instance status, health, billing history, plan info
- Rollback: redeploy previous version from deploy history
- Portal link: copy-to-clipboard button on instance detail page
- Portal layout: minimal, clean, no auth required
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

- Health check service pings instance URLs, shows status indicators
- Env var editor with add/remove/save (single CapRover API call)
- Activity timeline on instance detail page (from audit log)
- Retry deploy button for failed instances
- Maintenance mode toggle (scale to 0/1 replicas)
- Dark mode via Tailwind + localStorage toggle
- Dashboard: health check all, recent activity feed, total revenue
- Instance list: health dots, failed filter, responsive columns
- Logout button in header
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

getAppInfo only returns build status. getAppDefinition returns the full
config (env vars, volumes, replicas, SSL). Also auto-loads state on page
load so you don't have to click "Load State".
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Failed provisioning now appends '-failed-{timestamp}' to app_name,
freeing it for retry without manual cleanup.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

The forceSsl() update call only sends forceSsl:true, which causes
CapRover to reset all other fields including the git webhook config.
Wildcard cert + URL::forceScheme('https') handles HTTPS already.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Appends '-deleted-{timestamp}' to app_name when instance is deleted,
so the name can be reused for new instances.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

CapRover only allows lowercase letters and single hyphens. The generator
now strips invalid chars, forces lowercase suffix, and ignores deleted
instances when checking for collisions.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

The while/read loop was failing to properly quote env values with spaces
(APP_NAME="El Captain Manager" caused dotenv parse failure). sed handles
all edge cases (spaces, multiple = signs, special chars) in one pass.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

- Add configureAppFull() to CapRoverService that sends port, instanceCount,
  forceSsl, volumes, envVars, and appPushWebhook in ONE update call
  (CapRover wipes fields not included in each call)
- Rewrite InstanceProvisionerService::deployApp() to use configureAppFull()
- Add progress callback through provision() → deployApp() chain
- Add deploy progress panel in wizard (loading state + completion log)
- Fix provision() to not wrap in DB::transaction (was blocking progress)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

key:generate writes to .env, config:cache reads from .env — order matters.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Dumps runtime env vars into .env so config:cache picks them up
instead of reading stale .env.example defaults.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

config:cache bakes .env values at build time, ignoring Docker env vars
at runtime. Removed it so env vars from CapRover are read directly.
Fixed DB_HOST and CAPROVER_API_URL defaults in .env.example.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Replace broken deployFromGit() with setGitRepo() + forceBuild().
Add Force Build button per instance + Force Build All on list page.
No auto-deploy — builds only trigger when user clicks the button.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Laravel's key:generate reads the .env file content; it must exist and be
parseable before the command runs.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

The php:8.4-fpm-alpine image listens on TCP by default, not unix socket.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

CapRover injects env vars via Docker, not a .env file. The key:generate
command needs the file to exist to write the generated key.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Load State button fetches replicas, build status, SSL, env var count,
volumes, and recent deploy versions from CapRover API. Fetch Logs shows
the container's stdout/stderr in a dark terminal-style panel with refresh.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Instead of creating a separate PostgreSQL container per El Captain instance,
all instances now share the existing `elcaptainsportsonlydb` container.
Provisioner creates a dedicated database + user per instance via the
`pg_admin` superuser connection.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Full platform for operating El Captain as a service:
- Deploy new academies via one-click wizard (CapRover API integration)
- Flexible billing: monthly fee, platform %, or both per client
- Daily fee collection from instance databases (pull + webhook verify)
- Auto-suspend instances after grace period on overdue payments
- Dashboard with MRR, platform fees, instance health overview
- Complete audit log of all actions

Stack: Laravel + Livewire + Tailwind, deployed to CapRover
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>