Precision gets lost for big number in telegraf

Question

Precision gets lost for big number. I am using tail input plugin to read file and data inside a file is in json format. Below is the configuration

[inputs.tail]]
    files = ["E:/Telegraph/MSTCIVRRequestLog_*.json"]
    from_beginning = true
    name_override = "tcivrrequest"
    data_format = "json"
    json_strict = true

[[outputs.file]]
    files = ["E:/Telegraph/output.json"]
    data_format = "json"

Input file contains

{"RequestId":959011990586458245}

Expected Output

{"fields":{"RequestId":959011990586458245},"name":"tcivrrequest","tags":{},"timestamp":1632994599}

Actual Output

{"fields":{"RequestId":959011990586458200},"name":"tcivrrequest","tags":{},"timestamp":1632994599}

Number 959011990586458245 converted into 959011990586458200(check last few digits).

Already Tried Below things but not worked

json_string_fields = ["RequestId"]

[[processors.converter]] [processors.converter.fields] string = [""RequestId""]"

precision = "1s"

json_int64_fields = ["RequestId"]

character_encoding = "utf-8"

json_strict = true

Answer 1

I was able to reproduce this with the json parser as well. My suggestion would be to move to the json_v2 parser with a config like the following:

[[inputs.file]]
    files = ["metrics.json"]
    data_format = "json_v2"
    [[inputs.file.json_v2]]
        [[inputs.file.json_v2.field]]
        path = "RequestId"
        type = "int"

I was able to get a result as follows:

file RequestId=959011990586458245i 1651181595000000000

The newer parser is generally more accurate and flexible for simple cases like the one you provided.

Thanks!