IoT Telemetry Platform

A Laravel backend for solar/energy device monitoring: ingests telemetry from inverters and meters through two independent paths (a vendor cloud API and a direct on-site database poll), normalises it into hourly and daily energy rollups, raises priority-scored alerts, and turns unresolved alerts into tickets — with a full project/building/device hierarchy and per-resource permissions.

This is a generalised reference implementation of a production system. The real folder structure, ingestion pipelines, rollup strategy and alerting model are preserved; employer, vendor and site names are replaced with neutral ones, and the database dump plus every credential are removed.

What it demonstrates

• Two independent ingestion paths for the same data — a vendor cloud API (VendorApiDeviceSync, VendorApiHistoricalDataSync, VendorApiAlertSync) and a direct poll of an on-site database (HeliosDbPollCommand, HeliosMeterDbPollCommand). Sites differ in what they expose, and a cloud API that goes down must not stop data collection from a site that can be read directly.
• Backlog replay as a first-class operation — ReplayTelemetryFeedBacklog and ProcessTelemetryFeedJsonFilesCommand reprocess raw payloads that arrived while the pipeline was down. Device telemetry cannot be re-requested after the fact, so raw payloads are archived (ArchiveTelemetryFeedRawFiles) before parsing and can be replayed.
• Rollups are rebuildable, not append-only — RebuildDailyDeviceSummaries and TelemetryFeedRecomputeDailyEnergy recompute DailyDeviceSummary / DailyProjectSummary / HourlyDevicePowerGeneration from source records. Late-arriving or corrected telemetry is normal in device fleets, so a summary that can only be appended to is permanently wrong the first time a device backfills.
• Alert priority as configuration — AlertPriorityMapping + AlertPriorityMappingService, with BackfillAlertPriorities to reapply a changed mapping to alerts already raised.
• Alerts become tickets automatically — CreateMissingTickets reconciles alerts that should have produced a ticket and did not, so a failure in the alert→ticket path is self-healing rather than silently losing work.
• Live power cached separately from history — TelemetryFeedLivePowerCacheService serves dashboard "current output" without touching the historical tables the rollups are built from.
• Resource-scoped permissions — UserResourcePermission alongside Role/Permission, so access is granted per project/building/device rather than globally.
• Internal code generators — MakeCrudCommand and MakeServiceCommand with stubs/, keeping new resources structurally consistent.
• Error paths are recorded — DeviceGenerationErrorLog and TelemetryFeedPayloadLogger keep what failed to parse, so a malformed payload is diagnosable instead of just dropped.

Structure

app/
  Console/Commands/
    VendorApi{Device,Project,Alert,HistoricalData}Sync.php   cloud ingestion
    HeliosDbPollCommand.php · HeliosMeterDbPollCommand.php   on-site DB polling
    ProcessTelemetryFeedJsonFilesCommand.php                 raw payload parsing
    ReplayTelemetryFeedBacklog.php                           gap recovery
    ArchiveTelemetryFeedRawFiles.php                         raw retention
    RebuildDailyDeviceSummaries.php
    TelemetryFeedRecomputeDailyEnergy.php                    rollup rebuilds
    BackfillAlertPriorities.php · CreateMissingTickets.php   reconciliation
    MakeCrudCommand.php · MakeServiceCommand.php             scaffolding
  Jobs/                 vendor API fetch/parse/import, alert processing,
                        ticket notification
  Services/             Alert, AlertPriorityMapping, Device, Project, Building,
                        Entity, Tariff, Ticket, Dashboard, Permission, Role,
                        User, FCM, Sms, TelemetryFeedLivePowerCache,
                        TelemetryFeedPayloadLogger, Apis/
  Models/               Device · DeviceGeneration(+ErrorLog)
                        HourlyDevicePowerGeneration
                        DailyDeviceSummary · DailyProjectSummary
                        Project · Building · Entity · Tariff
                        Alert · AlertPriorityMapping
                        Ticket · TicketComment · TicketHistory
                        Role · Permission · UserResourcePermission
config/  database/  routes/  stubs/  supervisors/  documentation/  tests/

Running

cp .env.example .env      # DB, vendor API, on-site DB, FCM, SMS
composer install
php artisan migrate
php artisan schedule:work     # drives polling, rollups and reconciliation
php artisan queue:work

Ships without vendor/, without any database dump, and with every credential read from the environment.