veza/veza-backend-api/cmd/api/main.go
senke 9cdfc6d898 fix(backend): J4 — GDPR-compliant hard delete with Redis and ES cleanup
Closes TODO(HIGH-007). When the hard-delete worker anonymizes a user past
their recovery deadline, it now also cleans the user's residual data from
Redis and Elasticsearch, not just PostgreSQL. Without this, a user who
invoked their right to erasure would still appear in cached feed/profile
responses and in ES search results for up to the next reindex cycle.

Worker changes (internal/workers/hard_delete_worker.go):

  WithRedis / WithElasticsearch builder methods inject the clients. Both
  are optional: if either is nil (feature disabled or unreachable), the
  corresponding cleanup is skipped with a debug log and the worker keeps
  going. Partial progress beats panic.

  cleanRedisKeys uses SCAN with a cursor loop (COUNT 100), NEVER KEYS —
  KEYS would block the Redis server on multi-million-key deployments.
  Pattern is user:{id}:*. Transient SCAN errors retry up to 3 times with
  100ms * retry linear backoff; persistent errors return without panic.
  DEL errors on a batch are logged but non-fatal so subsequent batches
  are still attempted.

  cleanESDocs hits three indices independently:
    - users index: DELETE doc by _id (the user UUID); 404 treated as
      success (already gone = desired state)
    - tracks index: DeleteByQuery with a terms filter on _id, using the
      list of track IDs collected from PostgreSQL BEFORE anonymization
    - playlists index: same pattern as tracks
  A failure on one index does not prevent the others from being tried;
  the first error is returned so the caller can log.

  Track/playlist IDs are pre-collected (collectTrackIDs, collectPlaylistIDs)
  before the UPDATE anonymization runs, because the anonymization does NOT
  cascade (no DELETE on users), so tracks and playlists rows remain with
  their creator_id / user_id intact and resolvable at query time.

Wiring (cmd/api/main.go):

  The worker now receives cfg.RedisClient directly, and an optional ES
  client built from elasticsearch.LoadConfig() + NewClient. If ES is
  disabled or unreachable at startup, the worker logs a warning and
  proceeds with Redis-only cleanup.

Tests (internal/workers/hard_delete_worker_test.go, +260 lines):

  Pure-function unit tests:
    - TestUUIDsToStrings
    - TestEsIndexNameFor
  Nil-client safety tests:
    - TestCleanRedisKeys_NilClientIsNoop
    - TestCleanESDocs_NilClientIsNoop
  ES mock-server tests (httptest.Server mimicking /_doc and
  /_delete_by_query endpoints with valid ES 8.11 responses):
    - TestCleanESDocs_CallsAllThreeIndices — verifies the three expected
      HTTP calls land with the right paths and request bodies containing
      the provided UUIDs
    - TestCleanESDocs_SkipsEmptyIDLists — verifies no DeleteByQuery is
      issued when the ID lists are empty
  Redis testcontainer integration test (gated by VEZA_SKIP_INTEGRATION):
    - TestCleanRedisKeys_Integration — seeds 154 keys (4 fixed + 150 bulk
      to force the SCAN loop past a single batch) plus 4 unrelated keys
      from another user / global, runs cleanRedisKeys, asserts all 154
      own keys are gone and all 4 unrelated keys remain.

Verification:
  go build ./...                                                OK
  go vet ./...                                                  OK
  VEZA_SKIP_INTEGRATION=1 go test ./internal/workers/... short  OK
  go test ./internal/workers/ -run TestCleanRedisKeys_Integration
    → testcontainers spins redis:7-alpine, test passes in 1.34s

Out of J4 scope (noted for a follow-up):
  - No "activity" ES index exists in the codebase today (the audit plan
    mentioned it as a possible target). The three real indices with user
    data — users, tracks, playlists — are all now cleaned.
  - Track artist strings (free-form) may still contain the user's
    display name as a cached value in the tracks index after this
    cleanup. Actual user-owned tracks are deleted here, but if a third
    party's track referenced the removed user in its artist field, that
    reference is not touched. Strict RGPD on that edge case is a
    separate ticket.

Refs: AUDIT_REPORT.md §8.5, §10 P5, §12 item 1
2026-04-15 12:25:39 +02:00

343 lines
12 KiB
Go
Raw Blame History

This file contains invisible Unicode characters

This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

package main
import (
"context"
"fmt"
"log"
"net/http"
// SECURITY(REM-027): pprof removed from production — use build tag or dedicated debug binary instead.
// To enable: go build -tags debug ./cmd/api
"os"
"os/signal"
"syscall"
"time"
"github.com/getsentry/sentry-go"
"github.com/gin-gonic/gin"
"github.com/joho/godotenv"
"go.uber.org/zap"
"veza-backend-api/internal/api"
"veza-backend-api/internal/config"
"veza-backend-api/internal/core/marketplace"
vezaes "veza-backend-api/internal/elasticsearch"
"veza-backend-api/internal/metrics"
"veza-backend-api/internal/services"
"veza-backend-api/internal/shutdown"
"veza-backend-api/internal/workers"
_ "veza-backend-api/docs" // Import docs for swagger
)
// @title Veza Backend API
// @version 1.2.0
// @description Backend API for Veza platform.
// @termsOfService http://swagger.io/terms/
// @contact.name API Support
// @contact.url http://www.veza.app/support
// @contact.email support@veza.app
// @license.name Apache 2.0
// @license.url http://www.apache.org/licenses/LICENSE-2.0.html
// @host localhost:18080
// @BasePath /api/v1
// @securityDefinitions.apikey BearerAuth
// @in header
// @name Authorization
// @securityDefinitions.apikey ApiKeyAuth
// @in header
// @name X-API-Key
// @description Developer API key (obtain from Developer Portal). Format: vza_xxxxx
func main() {
// Charger les variables d'environnement
// NOTE: Do not write to stderr to avoid broken pipe errors with systemd journald
// The message will be logged by the logger once it's initialized
_ = godotenv.Load()
// FIX #1: Supprimer l'initialisation dupliquée du logger
// Le logger sera initialisé dans config.NewConfig() avec le bon LOG_LEVEL
// Charger la configuration (qui initialise le logger)
cfg, err := config.NewConfig()
if err != nil {
// CRITICAL: Do not write to stderr or files to avoid broken pipe errors
// Just exit silently - systemd will capture the exit code
// The error details will be in the application logs if the logger was initialized
os.Exit(1)
}
// Utiliser le logger de la config
logger := cfg.Logger
if logger == nil {
log.Fatal("❌ Logger non initialisé dans la configuration")
}
logger.Info("🚀 Démarrage de Veza Backend API")
// Valider la configuration
if err := cfg.Validate(); err != nil {
logger.Fatal("❌ Configuration invalide", zap.Error(err))
}
// Initialiser Sentry si DSN configuré
if cfg.SentryDsn != "" {
err := sentry.Init(sentry.ClientOptions{
Dsn: cfg.SentryDsn,
Environment: cfg.SentryEnvironment,
TracesSampleRate: cfg.SentrySampleRateTransactions,
SampleRate: cfg.SentrySampleRateErrors,
// AttachStacktrace pour capturer les stack traces
AttachStacktrace: true,
})
if err != nil {
logger.Warn("❌ Impossible d'initialiser Sentry", zap.Error(err))
} else {
logger.Info("✅ Sentry initialisé", zap.String("environment", cfg.SentryEnvironment))
}
// Flush les événements Sentry avant shutdown
defer sentry.Flush(2 * time.Second)
} else {
logger.Info(" Sentry non configuré (SENTRY_DSN non défini)")
}
// Initialisation de la base de données
db := cfg.Database
if db == nil {
logger.Fatal("❌ Base de données non initialisée")
}
defer db.Close()
if err := db.Initialize(); err != nil {
logger.Fatal("❌ Impossible d'initialiser la base de données", zap.Error(err))
}
// MOD-P2-004: Démarrer le collecteur de métriques DB pool
// Collecte les stats DB pool toutes les 10 secondes et les expose via Prometheus
metrics.StartDBPoolStatsCollector(db.DB, 10*time.Second)
logger.Info("✅ Collecteur de métriques DB pool démarré")
// Fail-Fast: Vérifier RabbitMQ si activé
if cfg.RabbitMQEnable {
if cfg.RabbitMQEventBus == nil {
logger.Fatal("❌ RabbitMQ activé (RABBITMQ_ENABLE=true) mais non initialisé (problème de connexion?)")
} else {
// Optionnel: Check connection status if RabbitMQEventBus exposes it
// For now, assume if initialized it's connected or retrying.
// If we want STRICT fail fast, we would need to verify connection is Open here.
logger.Info("✅ RabbitMQ actif")
}
} else {
logger.Info(" RabbitMQ désactivé")
}
// BE-SVC-017: Créer le gestionnaire de shutdown gracieux
shutdownManager := shutdown.NewShutdownManager(logger)
// Démarrer le Job Worker avec contexte pour shutdown gracieux
var workerCtx context.Context
var workerCancel context.CancelFunc
if cfg.JobWorker != nil {
workerCtx, workerCancel = context.WithCancel(context.Background())
cfg.JobWorker.Start(workerCtx)
logger.Info("✅ Job Worker démarré")
// Enregistrer le Job Worker pour shutdown gracieux
shutdownManager.Register(shutdown.NewShutdownFunc("job_worker", func(ctx context.Context) error {
if workerCancel != nil {
workerCancel()
// Attendre un peu pour que les workers se terminent
time.Sleep(2 * time.Second)
}
return nil
}))
} else {
logger.Warn("⚠️ Job Worker non initialisé")
}
// v0.701: Start Transfer Retry Worker
if cfg.TransferRetryEnabled && cfg.StripeConnectEnabled && cfg.StripeConnectSecretKey != "" {
stripeConnectSvc := services.NewStripeConnectService(db.GormDB, cfg.StripeConnectSecretKey, logger)
retryWorker := marketplace.NewTransferRetryWorker(
db.GormDB, stripeConnectSvc, logger, cfg.TransferRetryInterval, cfg.TransferRetryMaxAttempts,
)
retryCtx, retryCancel := context.WithCancel(context.Background())
go retryWorker.Start(retryCtx)
logger.Info("Transfer Retry Worker started",
zap.Duration("interval", cfg.TransferRetryInterval),
zap.Int("max_retries", cfg.TransferRetryMaxAttempts))
shutdownManager.Register(shutdown.NewShutdownFunc("transfer_retry_worker", func(ctx context.Context) error {
retryCancel()
return nil
}))
} else if cfg.TransferRetryEnabled {
logger.Info("Transfer Retry Worker skipped — Stripe Connect not enabled")
}
// v0.802: Start Cloud Backup Worker (copies cloud files to backup prefix every 24h)
if cfg.S3StorageService != nil {
backupWorker := services.NewCloudBackupWorker(db.GormDB, cfg.S3StorageService, logger)
backupCtx, backupCancel := context.WithCancel(context.Background())
go backupWorker.Start(backupCtx)
logger.Info("Cloud Backup Worker started (24h interval)")
shutdownManager.Register(shutdown.NewShutdownFunc("cloud_backup_worker", func(ctx context.Context) error {
backupCancel()
return nil
}))
}
// v0.802: Start Gear Warranty Notifier (sends notifications when warranty expires in 30 days)
notificationService := services.NewNotificationService(db, logger)
warrantyNotifier := services.NewGearWarrantyNotifier(db.GormDB, notificationService, logger)
warrantyCtx, warrantyCancel := context.WithCancel(context.Background())
go warrantyNotifier.Start(warrantyCtx)
logger.Info("Gear Warranty Notifier started (24h interval)")
shutdownManager.Register(shutdown.NewShutdownFunc("gear_warranty_notifier", func(ctx context.Context) error {
warrantyCancel()
return nil
}))
// v0.10.5 F552: Weekly notification digest (runs on Sunday)
if cfg.JobWorker != nil {
digestWorker := services.NewNotificationDigestWorker(db.GormDB, cfg.JobWorker, logger)
digestCtx, digestCancel := context.WithCancel(context.Background())
go digestWorker.Start(digestCtx)
logger.Info("Notification digest worker started (weekly on Sunday)")
shutdownManager.Register(shutdown.NewShutdownFunc("notification_digest_worker", func(ctx context.Context) error {
digestCancel()
return nil
}))
}
// v0.10.8 F065: Hard delete worker (GDPR - final anonymization after 30 days)
if os.Getenv("HARD_DELETE_CRON_ENABLED") != "false" {
// Optional ES client for the worker's RGPD cleanup (users/tracks/playlists indices).
// Non-fatal if ES is disabled or unreachable — the worker will just skip ES cleanup.
var hardDeleteESClient *vezaes.Client
if esCfg := vezaes.LoadConfig(); esCfg.Enabled {
if esc, esErr := vezaes.NewClient(esCfg, logger); esErr != nil {
logger.Warn("Elasticsearch unavailable for hard delete worker, ES cleanup disabled",
zap.Error(esErr))
} else {
hardDeleteESClient = esc
}
}
hardDeleteWorker := workers.NewHardDeleteWorker(db.GormDB, logger, 24*time.Hour).
WithRedis(cfg.RedisClient).
WithElasticsearch(hardDeleteESClient)
hardDeleteCtx, hardDeleteCancel := context.WithCancel(context.Background())
go hardDeleteWorker.Start(hardDeleteCtx)
logger.Info("Hard delete worker started (24h interval)",
zap.Bool("redis_cleanup", cfg.RedisClient != nil),
zap.Bool("es_cleanup", hardDeleteESClient != nil),
)
shutdownManager.Register(shutdown.NewShutdownFunc("hard_delete_worker", func(ctx context.Context) error {
hardDeleteWorker.Stop()
hardDeleteCancel()
return nil
}))
} else {
logger.Info("Hard delete worker disabled (HARD_DELETE_CRON_ENABLED=false)")
}
// Configuration du mode Gin
// Correction: Utilisation directe de la variable d'env car non exposée dans Config
appEnv := os.Getenv("APP_ENV")
if appEnv == "production" {
gin.SetMode(gin.ReleaseMode)
} else {
gin.SetMode(gin.DebugMode)
}
// Créer le router Gin
router := gin.New()
// SECURITY(HIGH-006): Restrict trusted proxies to prevent IP spoofing via X-Forwarded-For.
// Default: trust nothing (c.ClientIP() returns RemoteAddr only).
// Set TRUSTED_PROXIES="10.0.0.1,10.0.0.2" if behind a known reverse proxy/load balancer.
router.SetTrustedProxies(nil)
// Middleware globaux (Logger, Recovery) recommandés par ORIGIN
router.Use(gin.Logger(), gin.Recovery())
// Configuration des routes
apiRouter := api.NewAPIRouter(db, cfg) // Instantiate APIRouter
if err := apiRouter.Setup(router); err != nil {
logger.Error("Failed to setup API routes", zap.Error(err))
os.Exit(1)
}
// Configuration du serveur HTTP
port := fmt.Sprintf("%d", cfg.AppPort)
if cfg.AppPort == 0 {
port = "8080"
}
server := &http.Server{
Addr: fmt.Sprintf(":%s", port),
Handler: router,
ReadTimeout: 30 * time.Second, // Standards ORIGIN
WriteTimeout: 30 * time.Second,
}
// BE-SVC-017: Enregistrer tous les services pour shutdown gracieux
// Enregistrer le serveur HTTP
shutdownManager.Register(shutdown.NewShutdownFunc("http_server", func(ctx context.Context) error {
return server.Shutdown(ctx)
}))
// Enregistrer la configuration (ferme DB, Redis, RabbitMQ, etc.)
shutdownManager.Register(shutdown.NewShutdownFunc("config", func(ctx context.Context) error {
return cfg.Close()
}))
// Enregistrer le logger pour flush final
shutdownManager.Register(shutdown.NewShutdownFunc("logger", func(ctx context.Context) error {
if logger != nil {
return logger.Sync()
}
return nil
}))
// Enregistrer Sentry pour flush final
if cfg.SentryDsn != "" {
shutdownManager.Register(shutdown.NewShutdownFunc("sentry", func(ctx context.Context) error {
sentry.Flush(2 * time.Second)
return nil
}))
}
// Gestion de l'arrêt gracieux
quit := make(chan os.Signal, 1)
signal.Notify(quit, syscall.SIGINT, syscall.SIGTERM)
go func() {
logger.Info("🌐 Serveur HTTP démarré", zap.String("port", port))
if err := server.ListenAndServe(); err != nil && err != http.ErrServerClosed {
logger.Fatal("❌ Erreur du serveur HTTP", zap.Error(err))
}
}()
// Attendre le signal d'arrêt
<-quit
logger.Info("🔄 Signal d'arrêt reçu, démarrage du shutdown gracieux...")
// BE-SVC-017: Arrêt gracieux coordonné de tous les services
shutdownCtx, shutdownCancel := context.WithTimeout(context.Background(), 30*time.Second)
defer shutdownCancel()
if err := shutdownManager.Shutdown(shutdownCtx); err != nil {
logger.Error("❌ Erreur lors du shutdown gracieux", zap.Error(err))
} else {
logger.Info("✅ Shutdown gracieux terminé avec succès")
}
}