Use String#bytesize instead of String#length in #format_event

[yosangwon]

I found some issues which cause texts truncated on Datadog Event Stream when there are some characters outside ASCII range.

For example, a Korean letter(Hangul) takes 3 bytes in UTF-8 but current implementation considers it taking a byte. so every time a Korean letters there will be two bytes truncated.

text = "한글 한 자당 세 바이트abcdefghijklmnopqr"
text.length # => 31
text.bytesize # => 49

# remaining 18 bytes will be ignored in dd-agent (or datadog)
Datadog.dogstatsd.event("info", text, tags: ["test"], alert_type: "info")

Replacing String#length with String#bytesize would fix this issue.

[albertvaka]

Hi @devleoper 😄Thanks for your contribution!

I gave a look at the code that receives these packets and indeed, it is counting bytes instead of characters: https://github.com/DataDog/datadog-agent/blob/master/pkg/dogstatsd/parser.go#L173

Merging!