I've setup an S3 as a repository for my cluster and when I create a snapshot, it seems to work, though in the bucket the only think I see created is the meta files.
For example if I do this request:
PUT https://mycreds@my_ip:8443/_snapshot/elk_s3_repository/test_snapshot
{
  "indices": "logstash-2017.11.24",
  "ignore_unavailable": true,
  "include_global_state": false
}
The snapshot created will eventually has this status:
{
    "snapshots": [
        {
            "snapshot": "test_snapshot",
            "uuid": "tqOQLw94TsmWcPuWoCUC_A",
            "version_id": 5040099,
            "version": "5.4.0",
            "indices": [
                "logstash-2017.11.24"
            ],
            "state": "SUCCESS",
            "start_time": "2017-12-02T01:33:01.854Z",
            "start_time_in_millis": 1512178381854,
            "end_time": "2017-12-02T01:33:17.189Z",
            "end_time_in_millis": 1512178397189,
            "duration_in_millis": 15335,
            "failures": [],
            "shards": {
                "total": 5,
                "failed": 0,
                "successful": 5
            }
        }
    ]
}
But when I look in the S3 bucket, the only files created are:
snap-tqOQLw94TsmWcPuWoCUC_A.dat .  13.5KB
snap-VI2VFbM5QYOUohR9ZQXCbQ.dat . 246B
meta-I2IjYzR-Tk6YEw-JhbQ-ZQ.dat       13.5K
snap-I2IjYzR-Tk6YEw-JhbQ-ZQ.dat .      246B
I would expect to see much larger files for a 59MB index.. What am I missing?