Laravel Task Orchestrator Laravel Package

Changed

• Simplified scheduled-run behavior: when a scheduled task cannot be started, Task Orchestrator now records a failed scheduled run with a clear failure reason.
• Scheduled runs that fail to start are not retried immediately and are attempted again only at the next regular cron due time.

Removed

• Removed missed scheduled-run monitoring command (task-orchestrator:monitor-scheduled-runs).
• Removed missed-run health escalation and dedicated missed/recovered scheduled-run email notifications.
• Removed scheduled monitoring configuration and related documentation sections.

Fixed

• Corrected synchronous mail dispatch for missed-run alert and recovery emails to ensure notifications reach recipients even when the queue worker is temporarily unavailable.

Changed

• Dropped Laravel 11 from the supported versions: the entire 11.x line currently carries unpatched security advisories. The package now officially requires Laravel 12 or 13.
• Updated illuminate/support constraint from ^11.0|^12.0|^13.0 to ^12.0|^13.0.
• CI test matrix narrowed to Laravel 12 and 13; removed advisory-policy workaround that was needed to force-resolve the blocked 11.x line.
• Updated README badge and requirements section to reflect Laravel 12 and 13 only.

Added

• Missed scheduled-run monitoring for all cron-based tasks, including health payload reporting and a dedicated task-orchestrator:monitor-scheduled-runs command.
• Scheduled-run alert emails with incident deduplication and recovery notifications.
• Monitoring configuration for grace period, command failure behavior and notification recipients.

Changed

• Scheduled task health now reports missed scheduled runs as critical so silent schedule gaps become visible immediately.
• Project configuration now uses a more tolerant queue-worker heartbeat threshold to reduce false skips during short restarts.
• Documentation was updated to describe the new scheduled monitoring and email notification behavior.

Fixed

• Prevented missed scheduled runs from remaining green in the dashboard when no failed job record was created.

Fixed

• Fixed GitHub Actions dependency resolution for the Laravel 11 test job by mapping it to orchestra/testbench:^9.15.
• Updated workflow composer commands to keep Testbench in require-dev and allow transitive dependency adjustments during matrix resolution.

Added

• Laravel 13 compatibility in package constraints (illuminate/support) and dev tooling (orchestra/testbench).
• GitHub Actions test matrix coverage for Laravel 13 (alongside Laravel 11 and 12).

Changed

• Updated README support matrix and badge to include Laravel 13.

Added

• Comprehensive PHPUnit test coverage across HTTP controllers, integration flows, console commands, queued jobs, support services, domain objects, registry behavior, and property-based invariants.
• GitHub Actions workflow for running the package PHPUnit test suite.
• Operations documentation covering scheduler setup, queue workers, health monitoring, stale run recovery, and common troubleshooting.

Changed

• Reworked the README for public package usage with clearer positioning, installation steps, quick-start discovery configuration, generic dependency pipeline examples, dashboard access, health checks, stale run recovery, screenshots, support status, and troubleshooting.
• Improved composer.json package metadata with a stronger description, Packagist keywords, homepage, and support links.
• Updated installation and authorization documentation to match the current package configuration and scheduler behavior.
• Updated README badges for Packagist version, downloads, license, PHP version, Laravel support, and test workflow status.

Added

• Task mail notifications: Configurable email notifications on task failure and recovery (success after previous failure).
• Global notification configuration via task-orchestrator.notifications (enabled, recipients).
• Per-task notification configuration via discovery metadata notifications key, with precedence over global settings.
• NotificationEvaluationAction orchestrates notification dispatch after terminal status transitions.
• NotificationConfigResolver resolves effective enabled state and recipient list with global/per-task precedence.
• RecoveryDetector determines if a successful run represents recovery from a previous failure.
• TaskFailedMailable and TaskRecoveredMailable queued mailables with Blade email templates.
• NotificationPayload value object for resolved notification data (validated emails, max 50 recipients).
• TaskDefinition::notifications() fluent method for immutable per-task notification config.
• DiscoveredTaskDefinitionFactory maps notifications key from discovery metadata.
• Cache-based idempotence to prevent duplicate notifications per run/status.
• Graceful degradation: notifications skip silently when illuminate/mail is not registered.
• RFC 5322 email validation with invalid address filtering and warning logs.
• Published Blade views via existing task-orchestrator-views publish tag for template customization.
• Property-based tests (innmind/black-box) covering config resolution, email filtering, recovery detection, idempotence, and mail exception resilience.
• Integration tests for failure notification flow, recovery notification flow, and service provider conditional registration.
• Unit tests for TaskFailedMailable and TaskRecoveredMailable.

Changed

• ExecuteTaskRunAction now accepts an optional NotificationEvaluationAction dependency and triggers notification evaluation after terminal status updates.
• ExecuteTaskRunJob::failed() now triggers notification evaluation for timeout/exception failures.
• TaskOrchestratorServiceProvider conditionally registers notification services when mail.manager is bound.

Fixed

• Queue flood prevention: QueueHeartbeatJob now implements ShouldBeUnique with a 120-second lock to prevent thousands of duplicate jobs accumulating when the queue worker is down.
• Scheduled task guard: RunScheduledTaskCommand now checks queue worker heartbeat health before dispatching. When the queue worker is down, scheduled tasks are skipped with a warning instead of creating orphaned task run records.
• Run history ordering: Replaced latest('started_at'), latest('created_at') with COALESCE(finished_at, started_at, created_at) DESC across all run queries. Runs with NULL started_at (stale queued runs marked as failed) now sort correctly by their most recent activity timestamp.

Changed

• DashboardController, DashboardStatusController, TaskRunIndexController, FailedTaskRunIndexController, TaskDetailDataProvider, and TaskStartBlockingEvaluator all use consistent COALESCE-based ordering for task run queries.

🎨 Enhanced

• UI/UX Consistency Overhaul: Complete visual redesign for consistent user experience across all views
• Status Badges: Added thematic icons to all status badges (succeeded ✅, failed ❌, running ▶️, queued ⏰, cancelled ⛔)
• Trigger Badges: Enhanced trigger badges with contextual icons (pipeline 🔗, scheduled ⏰, manual 🔒, retry 🔄)
• Group Icons: Intelligent thematic icons for task groups based on naming patterns (ETL 📥, Users 👥, Security 🔒, etc.)
• Duration Formatting: Smart duration display (45s, 22m 15s, 2h 15m) replacing raw seconds for better readability
• UUID Optimization: Shortened UUIDs to 6-8 characters across all tables and cards for cleaner interface
• Health Animation: Added subtle breathing animation to dashboard health icon for enhanced visual feedback
• Button Consistency: Unified button styling with proper icons across Run Detail and Task Detail views

🔧 Improved

• Badge Styling: Consistent flexbox layout, unified heights, and proper icon-text alignment
• CSS Architecture: Enhanced badge and pill classes with better symmetry and spacing
• Navigation Flow: Improved button placement and icon consistency in detail views
• Visual Hierarchy: Better contrast and spacing for improved readability

🎯 Technical

• Icon System: Comprehensive SVG icon library with consistent sizing (10px-14px) and stroke weights
• CSS Classes: Migrated from status-badge to status-pill for better consistency
• Component Harmony: Unified design patterns across Vue components and Blade templates
• Responsive Design: Maintained mobile-friendly layouts while enhancing desktop experience

Added

• Persistent per-run timeout tracking on task_runs via timeout_seconds and timeout_at so stale-run handling no longer depends solely on the current task definition.
• Regression coverage for task dispatch timeout persistence, stale-run recovery, and preserving original run deadlines across repeated processing attempts.

Changed

• StartTaskAction now stores the effective timeout on the task run record when dispatching ExecuteTaskRunJob.
• ExecuteTaskRunAction now preserves an existing run's original started_at and timeout_at values instead of resetting them on a later execution attempt.
• RecoverStaleTaskRunsCommand now resolves stale runs primarily from persisted per-run timeout data, while still supporting explicit global timeout overrides.

Fixed

• Fixed stale task runs remaining stuck in running after their configured timeout when worker timeout failure hooks were not reached reliably.
• Fixed recovery behavior for long-running tasks whose timeout definition changed after the run had already been queued or started.

Added

• Topbar redesign: logo icon, sticky positioning, active-state navigation indicators, user avatar with initials, and dedicated theme toggle with sun/moon SVG icons.
• Dashboard stat cards with accent-colored left borders, icons, and hover elevation effects.
• Health monitoring row: large status indicator with icon, status pills for queue/scheduler/worker, and metric cards with animated progress bars.
• Task group panels with color-coded group icons, inline run-history dots, dependency badges, schedule/trigger tags, and live progress bars for running tasks.
• Recent runs and failed runs panels with structured table layout and failure detail display.
• Task detail page: stat cards row (group, queue, timeout, last status, last run), info cards for dependencies and run history, and polished tab bar with pagination.
• Pagination component with styled page links, ellipsis, and navigation buttons.

Changed

• Full CSS rewrite from basic utility styles to a comprehensive design-system approach with design tokens, consistent spacing, shadows, and border-radius variables.
• Dashboard layout restructured from simple card grid to a rich multi-section dashboard (stat cards → health row → grouped tasks → recent/failed runs).
• Theme toggle improved: respects prefers-color-scheme on first visit, persists choice in localStorage, and toggles sun/moon icons instead of emoji.
• Topbar navigation now highlights the active route with an accent-colored underline indicator.
• Task detail page tabs bar: increased gap between tabs and added bottom padding so tabs no longer sit on the border.
• Task detail page tab pagination: buttons now have proper spacing (1rem gap) with a subtle top border separator.

Fixed

• Tabs in .td-tabs-bar had no bottom spacing causing them to overlap the border line.
• Pagination buttons in .task-tab-pagination were missing styles entirely, causing them to render without any gap.

Fixed

• Removed overlap protection from the scheduled task-orchestrator:record-scheduler-heartbeat command so scheduler liveness heartbeats are not blocked by stale scheduler mutexes in Cloud Foundry / distributed environments.

Added

• Queue worker liveness monitoring in dashboard health payload:
  • queue_worker.status (running, down)
  • queue_worker.last_heartbeat_at
• Queue worker heartbeat mechanisms:
  • scheduled heartbeat via QueueHeartbeatJob
  • execution heartbeat updates in ExecuteTaskRunJob
• Task detail page with lazy-loaded tabs and dedicated endpoints:
  • GET /tasks/{task}
  • GET /tasks/{task}/runs
  • GET /tasks/{task}/failures
  • GET /tasks/{task}/logs
  • GET /tasks/{task}/documentation
• TaskDetailDataProvider for task meta, recent runs, startability state, logs, and documentation extraction.
• Task documentation extraction support from command description, documentation() method, and $documentation property.
• Dashboard task labels now link directly to task detail pages.

Changed

• Queue worker heartbeat is now hybrid (scheduled + task execution) so liveness detection also works when scheduler is unavailable but workers process tasks.
• Manual task start now starts only the selected task (no upstream chain resolution).
• Task detail page UX polish:
  • structured overview cards/subsections
  • stronger tab styling and spacing
  • default tab is now Runs (redundant Overview tab removed)
  • logs tab simplified to latest-run logs only
  • task start action integrated into detail header using existing backend startability rules
• Removed redundant dashboard header action buttons (View tasks, View runs, View pipelines) because top navigation already covers these links.

Docs

• Updated docs with queue worker liveness behavior and health.queue_worker.heartbeat_max_age_seconds configuration.
• Updated documentation for task detail page behavior and manual start semantics.

Added

• Centralized task start blocking evaluator (TaskStartBlockingEvaluator) reused by UI state and backend guard logic.
• Dependency-aware startability metadata in task payloads:
  • is_blocked_by_dependencies
  • blocked_by_task_names
  • start_block_reason
  • can_start
• Dashboard health monitoring for queue and scheduler:
  • queue status (healthy, busy, stuck) based on pending job count and oldest pending job age
  • scheduler status (running, down) based on scheduler heartbeat timestamp
  • configurable health thresholds and heartbeat settings in task-orchestrator.health
• Scheduler heartbeat command and schedule registration:
  • task-orchestrator:record-scheduler-heartbeat
  • heartbeat recorded every minute via scheduler registration

Changed

• StartTaskAction now enforces dependency-based blocking through the centralized backend guard path.
• Stale-run recovery scheduler registration now runs without hardcoded --minutes override to respect task-specific timeouts.
• Manual task start now starts only the selected task (no upstream chain resolution).
• Dashboard health payload now includes structured queue and scheduler sections while keeping compatibility top-level metrics.

Fixed

• Prevented duplicate task starts across trigger paths when dependencies are active or not in a succeeded state.
• Disabled run buttons consistently when backend startability state is false in task list and dashboard views.
• Removed duplicate or unclear run-state helper actions in the UI and simplified button states.
• Replaced mixed pagination rendering with a single package-styled pagination component.
• Prevented manual starts of dependent tasks from automatically dispatching prerequisite tasks.

Docs

• Updated health monitoring docs to describe queue/scheduler status semantics and configuration.
• Updated pipeline behavior docs to reflect manual start behavior.

Added

• Per-task queue routing via discovery metadata:
  • queue
  • connection

Fixed

• TaskDefinition immutable builders now preserve queue and connection across chained calls.
• Retry no longer fails with runtime error when the same task is already queued/running; existing active run is returned instead.

Docs

• Updated README.md and docs/discovery.md with queue/connection discovery examples.

🚀 Added

• Task dependency execution (depends_on)
• Automatic pipeline execution (downstream triggering)
• Pipeline view UI
• Trigger types (manual, scheduled, pipeline, retry)
• Dark / Light mode toggle
• Responsive dashboard improvements
• Stale run recovery command
• Per-task timeout configuration (timeout_minutes)
• Discovery config inside application (discovery_path)

✨ Improved

• Dashboard UI redesign (cards, spacing, layout)
• Tables optimized for smaller screens
• Status badges improved
• Buttons and interactions polished
• Pipeline visualization UX

🛠 Fixed

• Trigger type flickering
• Duplicate task execution edge cases
• Running state inconsistencies
• Missing dependency handling