Activity

Summary

• Surfaced during PR #146 smoke testing on dev.tidylinker.com: queued jobs (DirectBookingRequestNotification, LogUserAuthActivity, NewSupportRequestNotification) all sat in Redis forever
• Root cause: config/horizon.php only defined production and local environments, but staging runs APP_ENV=staging, so Horizon spawned zero supervisors
• Adds a staging environment block (3 max processes — modest, matches the lighter staging traffic)
• Adds a Pest test asserting every env we actually run in has a supervisor block, plus a sanity check that horizon:list runs cleanly under APP_ENV=staging

Production was unaffected by the original bug.

Test plan

• php artisan test tests/Feature/Config/HorizonEnvironmentsTest.php → 4 passed
• Pint clean
• After merge + redeploy of dev.tidylinker.com: confirm php artisan horizon:list shows a running supervisor, and queued notifications drain automatically (previously had to drain manually with queue:work --once)