Parser plugin multiline doesn't work as it appears in docs

[satrushn]

Parser plugin multiline doesn't work as it appears in docs or there is a misunderstanding.
I tried to use parser multiline in section as it appears in docs : https://docs.fluentd.org/parser/multiline
I need to collect several lines in one message.

Example of log:

[DockerLogGenerator] Multiline: 2021-07-01 12:29:42.862326529 +0000 UTC m=+107095.440406775
This is the second line
This is the third line

I expect that this log will be parsed as something like this:

record:
{
"message":"[DockerLogGenerator] Multiline: 2021-07-01 12:29:42.862326529 +0000 UTC m=+107095.440406775\n This is the second line\n This is the third line
}

But it doesn't appears.

Example of my config:

Would you make clear in docs, how to collect multiline logs correctly in this case, please?
Thanks.

[kenhys]

I guess that <parse> in <source> may be what you want to do.

Here is the sample.

<source>
  @type tail
  path logs.txt
  tag test
  read_from_head true
  <parse>
     @type multiline
     format_firstline /\[/
     format1 /^(?<message>.*)/
  </parse>
</source>

<match test>
  @type stdout
</match>

OUTPUT:

2021-07-02 11:27:24 +0900 [info]: #0 following tail of logs.txt
2021-07-02 11:27:24.876553899 +0900 test: {"message":"[DockerLogGenerator] Multiline: 2021-07-01 12:29:42.862326529 +0000 UTC m=+107095.440406775\r\nThis is the second line\r\nThis is the third line"}
2021-07-02 11:27:24 +0900 [info]: #0 fluentd worker is now running worker=0

or it seems that there is a straightforward way copying from log to message by using record_transformer
(in above regex usage)

[satrushn]

Thank you for your answer.
I try to use Fluentd as Fluentd-operator, there is no way to change something in "source" that's why I want to use parser in filter.
Is there another method to solve the problem?

[kenhys]

I try to use Fluentd as Fluentd-operator, there is no way to change something in "source" that's why I want to use parser in filter.
Is there another method to solve the problem?

I do not understand well, but https://github.com/fluent-plugins-nursery/fluent-plugin-concat may help you.

[joek-office]

Hello together,
have seen this old post and it looks like that i have the same or a near problem. Hope anyone can clarify the documentation or give a way to solve my problem.

What i want to do:
I have many different files with different log formats. For this i have written a new regex to use with multiline. The jumping point here is, that i have many one line logs and multiline logs. i want to solve this like in the documentation (Java Stacktrace Log
). The problem is, that in the environment comes for multiline log output the following error message:
#0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log
and then follows the lines one by one from the log output.

'2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "[2023-07-10 14:04:31 +0200].535 35 WARNING [-] Invalid HTTP request received.\n"
2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "Traceback (most recent call last):\n"
2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "  File \"/var/lib/kolla/venv/lib64/python3.9/site-packages/uvicorn/protocols/http/h11_impl.py\", line 129, in handle_events\n"
2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "    event = self.conn.next_event()\n"
2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "  File \"/var/lib/kolla/venv/lib64/python3.9/site-packages/h11/_connection.py\", line 443, in next_event\n"
2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "    exc._reraise_as_remote_protocol_error()\n"
2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "  File \"/var/lib/kolla/venv/lib64/python3.9/site-packages/h11/_util.py\", line 76, in _reraise_as_remote_protocol_error\n"
2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "    raise self\n"
2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "  File \"/var/lib/kolla/venv/lib64/python3.9/site-packages/h11/_connection.py\", line 425, in next_event\n"
2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "    event = self._extract_next_receive_event()\n"
2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "  File \"/var/lib/kolla/venv/lib64/python3.9/site-packages/h11/_connection.py\", line 367, in _extract_next_receive_event\n"
2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "    event = self._reader(self._receive_buffer)\n"
2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "  File \"/var/lib/kolla/venv/lib64/python3.9/site-packages/h11/_readers.py\", line 68, in maybe_read_from_IDLE_client\n"
2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "    raise LocalProtocolError(\"illegal request line\")\n"
2023-07-10 14:04:31 +0200 [warn]: #0 got incomplete line before first line from /var/log/kolla/skyline/skyline-error.log: "h11._util.RemoteProtocolError: illegal request line\n"'

My configuration is as follows:

\<source>
  \@ type tail
  path /var/log/kolla/*/*-access.log,/var/log/kolla/*/*-error.log,/var/log/kolla/*/*_access.log,/var/log/kolla/*/*_error.log
  pos_file /var/run/td-agent/kolla-openstack-wsgi.pos
  tag kolla.*
  enable_watch_timer false
  \<parse>
    \@ type multiline
    format_firstline /^([^\/\r\nA-Z]*?\[|\[?)(?<Timestamp>((\+)?((?<=[A-Za-z]{4})\d{2,6}|(?<![A-Za-z]{4})(\d{2,6}|\w{3}))(\/|-|:|\s|\.)?){3,7}(?=\d{4}\])\d{4})\]?/
    format1 /^([^\/\r\nA-Z]*?\[|\[?)(?<Timestamp>((\+)?((?<=[A-Za-z]{4})\d{2,6}|(?<![A-Za-z]{4})(\d{2,6}|\w{3}))(\/|-|:|\s|\.)?){3,7}(?=\d{4}\])\d{4})\]?(?<Payload>.*)/
  \</parse>
\</source>

I match the following log formats with this regex:

'10.1.101.12 - - [28/Jun/2023:08:11:12 +0200]
[Tue Jun 27 12:07:40.882366 2023]
[2023-06-28 08:36:26 +0200].455
2023/06/28 08:36:20
2023-06-27 09:49:10.619993'
the regex looks for pieces of the strings, not for the whole strings
piece: [prefix][infix][suffix] up to seven times
prefix: could be a + sign, not mandatory
infix: 2-6 digits or 3 word characters
suffix: one of /,-,:,. or space char

after the up to seven pieces it is possible to have a additional four digits if no characters are present

Can anyone help or advise why fluentd cant parse this log output as multiline and give the error message?