milliseconds accuracy about logback-gelf HOT 3 CLOSED

This library uses timestamps with millisecond precision.

I suspect a misconfiguration in your Graylog installation. As I don't see it in your screenshot – which field are you displaying? Is it the 'timestamp' field? Have you configured a pipeline rule which might cause this effect?

Could you try curl to debug this a bit further? You can find examples in the Graylog documentation about GELF.

from logback-gelf.

I posted the same question on graylog forums at this post.

They are telling that the problem is on elasticsearch...
Do you really think is that?

According to your questions:

1. Yes, the field displayed is timestamp.
2. The only pipeline rule I have is:

rule "uploaded size"
when
  starts_with(to_string($message.message), "Verificant si hi ha", true)
then
  let extract = grok("%{BASE10NUM:uploaded_size;long}", replace(to_string($message.message), "."), false);
  set_fields(extract);
end

3. I've tested with: echo -n -e '{ "version": "1.1", "host": "example.org", "short_message": "A short message", "level": 5, "_some_info": "foo" }'"\0" | nc -w0 localhost 12201
   The problem remains:
   

from logback-gelf.

I don’t think, this issue is related to elasticsearch and ES is definitely capable of storing timestamps with millisecond precision (see Date datatype).

As you can see in your own (latest) screenshot, the milliseconds are recorded. There are three timestamps displayed in the screenshot – two are missing the milliseconds and one is showing them. If you click on the UUID you will probably see even more fields with millisecond precision.

I tried to reproduce the problem and was able to reproduce it with Graylog 3.2.3. In Graylog 3.1.4 milliseconds are shown everywhere.

I hope I could help a little with this, but I close this ticket now as this is not a logback-gelf issue at all.

from logback-gelf.