PostgreSQL を用いた Go 言語でのトランザクショナルアウトボックスパターン

BEGIN;
  INSERT INTO orders ...   -- 成功
COMMIT;

PUBLISH order.created ...  -- 失敗、クラッシュ、または到達しない

BEGIN;
  INSERT INTO orders ...         -- ビジネスデータ
  INSERT INTO outbox_events ...  -- イベントレコード
COMMIT;

-- リレープロセス（別個に）:
SELECT ... FROM outbox_events FOR UPDATE SKIP LOCKED;
PUBLISH order.created ...
UPDATE outbox_events SET processed_at = NOW() WHERE id = $1;

CREATE TABLE outbox_events (
    id             UUID         PRIMARY KEY DEFAULT gen_random_uuid(),
    aggregate_type VARCHAR(100) NOT NULL,
    aggregate_id   VARCHAR(100) NOT NULL,
    event_type     VARCHAR(100) NOT NULL,
    payload        JSONB        NOT NULL,
    attempts       INT          NOT NULL DEFAULT 0,
    created_at     TIMESTAMPTZ  NOT NULL DEFAULT NOW(),
    processed_at   TIMESTAMPTZ
);

-- 部分インデックス：未処理の行のみをインデックス化し、行が完了としてマークされるにつれて小さく保つ
CREATE INDEX idx_outbox_unprocessed
    ON outbox_events (created_at)
    WHERE processed_at IS NULL;

type OrderService struct {
    db *sql.DB
}

func (s *OrderService) CreateOrder(ctx context.Context, order Order) error {
    tx, err := s.db.BeginTx(ctx, nil)
    if err != nil {
        return fmt.Errorf("begin tx: %w", err)
    }
    defer tx.Rollback()

    if _, err := tx.ExecContext(ctx, `
        INSERT INTO orders (id, customer_id, total, created_at)
        VALUES ($1, $2, $3, NOW())
    `, order.ID, order.CustomerID, order.Total); err != nil {
        return fmt.Errorf("insert order: %w", err)
    }

    payload, err := json.Marshal(map[string]any{
        "order_id":    order.ID,
        "customer_id": order.CustomerID,
        "total":       order.Total,
    })
    if err != nil {
        return fmt.Errorf("marshal payload: %w", err)
    }

    if _, err := tx.ExecContext(ctx, `
        INSERT INTO outbox_events
            (aggregate_type, aggregate_id, event_type, payload)
        VALUES ($1, $2, $3, $4)
    `, "order", order.ID, "order.created", payload); err != nil {
        return fmt.Errorf("insert outbox event: %w", err)
    }

    return tx.Commit()
}

type OutboxRelay struct {
    db          *sql.DB
    publisher   Publisher
    logger      *slog.Logger
    batchSize   int
    pollInterval time.Duration
    maxAttempts  int
}

func (r *OutboxRelay) Run(ctx context.Context) error {
    ticker := time.NewTicker(r.pollInterval)
    defer ticker.Stop()

    for {
        select {
        case <-ctx.Done():
            return ctx.Err()
        case <-ticker.C:
            if err := r.processBatch(ctx); err != nil {
                r.logger.Error("outbox relay batch failed", "err", err)
            }
        }
    }
}

func (r *OutboxRelay) processBatch(ctx context.Context) error {
    tx, err := r.db.BeginTx(ctx, nil)
    if err != nil {
        return fmt.Errorf("begin tx: %w", err)
    }
    defer tx.Rollback()

    rows, err := tx.QueryContext(ctx, `
        SELECT id, aggregate_type, aggregate_id, event_type, payload
        FROM outbox_events
        WHERE processed_at IS NULL
          AND attempts < $1
        ORDER BY created_at
        LIMIT $2
        FOR UPDATE SKIP LOCKED
    `, r.maxAttempts, r.batchSize)
    if err != nil {
        return fmt.Errorf("query outbox: %w", err)
    }
    defer rows.Close()

    type row struct {
        id            string
        aggregateType string
        aggregateID   string
        eventType     string
        payload       json.RawMessage
    }

    var batch []row
    for rows.Next() {
        var e row
        if err := rows.Scan(
            &e.id, &e.aggregateType, &e.aggregateID, &e.eventType, &e.payload,
        ); err != nil {
            return fmt.Errorf("scan row: %w", err)
        }
        batch = append(batch, e)
    }
    if err := rows.Err(); err != nil {
        return err
    }

    for _, e := range batch {
        if err := r.publisher.Publish(ctx, e.eventType, e.aggregateID, e.payload); err != nil {
            r.logger.Error("publish failed", "event_id", e.id, "err", err)
            if _, err := tx.ExecContext(ctx,
                `UPDATE outbox_events SET attempts = attempts + 1 WHERE id = $1`, e.id,
            ); err != nil {
                r.logger.Error("increment attempts failed", "event_id", e.id, "err", err)
            }
            continue
        }

        if _, err := tx.ExecContext(ctx,
            `UPDATE outbox_events SET processed_at = NOW() WHERE id = $1`, e.id,
        ); err != nil {
            return fmt.Errorf("mark processed: %w", err)
        }
    }

    return tx.Commit()
}

CREATE TABLE processed_events (
    event_id   UUID PRIMARY KEY,
    processed_at TIMESTAMPTZ NOT NULL DEFAULT NOW()
);

func (h *OrderHandler) HandleOrderCreated(ctx context.Context, eventID string, payload []byte) error {
    // イベントIDを自然キーとして使用して重複を排除
    _, err := h.db.ExecContext(ctx, `
        INSERT INTO processed_events (event_id) VALUES ($1)
        ON CONFLICT (event_id) DO NOTHING
    `, eventID)
    if err != nil {
        return fmt.Errorf("dedup check: %w", err)
    }

    // インサートが実際に発生したか（1行）または何もしなかったか（0行）を確認
    // よりシンプルなアプローチ：RETURNINGを使用するか、影響を受けた行をチェック
    // 0行が影響を受けた場合、これは重複です -- スキップ
    ...
}

CREATE VIEW outbox_dead_letters AS
SELECT *
FROM outbox_events
WHERE attempts >= 5
  AND processed_at IS NULL
ORDER BY created_at;

ORDER BY aggregate_id, created_at

CREATE OR REPLACE FUNCTION notify_outbox_insert() RETURNS trigger AS $$
BEGIN
    PERFORM pg_notify('outbox_event', NEW.id::text);
    RETURN NEW;
END;
$$ LANGUAGE plpgsql;

CREATE TRIGGER outbox_insert_notify
AFTER INSERT ON outbox_events
FOR EACH ROW EXECUTE FUNCTION notify_outbox_insert();

func (r *OutboxRelay) Run(ctx context.Context) error {
    listener := pq.NewListener(r.dsn, 10*time.Second, time.Minute, nil)
    defer listener.Close()

    if err := listener.Listen("outbox_event"); err != nil {
        return fmt.Errorf("listen: %w", err)
    }

    ticker := time.NewTicker(5 * time.Second) // fallback poll
    defer ticker.Stop()

    for {
        select {
        case <-ctx.Done():
            return ctx.Err()
        case <-listener.Notify:
            if err := r.processBatch(ctx); err != nil {
                r.logger.Error("outbox batch failed (notify)", "err", err)
            }
        case <-ticker.C:
            if err := r.processBatch(ctx); err != nil {
                r.logger.Error("outbox batch failed (poll)", "err", err)
            }
        }
    }
}

var (
    outboxPublished = prometheus.NewCounter(prometheus.CounterOpts{
        Name: "outbox_events_published_total",
        Help: "Total outbox events successfully published.",
    })
    outboxFailed = prometheus.NewCounterVec(prometheus.CounterOpts{
        Name: "outbox_events_failed_total",
        Help: "Total outbox publish failures by event type.",
    }, []string{"event_type"})
    outboxPending = prometheus.NewGauge(prometheus.GaugeOpts{
        Name: "outbox_events_pending",
        Help: "Current number of unprocessed outbox events.",
    })
    outboxBatchDuration = prometheus.NewHistogram(prometheus.HistogramOpts{
        Name:    "outbox_batch_duration_seconds",
        Help:    "Duration of each outbox processing batch.",
        Buckets: prometheus.DefBuckets,
    })
)

SELECT COUNT(*) FROM outbox_events WHERE processed_at IS NULL;

-- name: InsertOutboxEvent :exec
INSERT INTO outbox_events (aggregate_type, aggregate_id, event_type, payload)
VALUES (@aggregate_type, @aggregate_id, @event_type, @payload);

-- name: FetchOutboxBatch :many
SELECT id, aggregate_type, aggregate_id, event_type, payload
FROM outbox_events
WHERE processed_at IS NULL
  AND attempts < @max_attempts
ORDER BY created_at
LIMIT @batch_size
FOR UPDATE SKIP LOCKED;

-- name: MarkOutboxProcessed :exec
UPDATE outbox_events SET processed_at = NOW() WHERE id = @id;

-- name: IncrementOutboxAttempts :exec
UPDATE outbox_events SET attempts = attempts + 1 WHERE id = @id;

-- name: OutboxPendingCount :one
SELECT COUNT(*) FROM outbox_events WHERE processed_at IS NULL;