AWS Glue JSON to Parquet conversion

Question

I am trying to convert JSON files into Parquet using AWS Glue containing data formatted like this:

[
  { "id": 1, "message": "test message of event 1" },
  { "id": 2, "message": "test message of event 2" },
  { "id": 3, "message": "test message of event 3" }
]

After setting up all of the options (source bucket folder, target bucket folder), I get this output schema:

Data format is set up to JSON, with no JsonPath (it doesn't allow me to specify it), selected multiline JSON then I clicked Infer Schema.

As a result, first row from the array is always omitted:

Does anybody know why Glue omits the first record while doing the conversion?