Feature request: Store raw data before parsing

[abangser]

Thanks so much for making this parser, it is exactly what I was looking for. My only additional hope is to retain the audit log in its raw form as well.

I have tried to chain this filter together with another filter like this:

<source>
  @type tail
  format none
  path /var/log/audit/audit.log
  read_from_head false
  tag audit
</source>

<filter audit>
  @type record_transformer
  <record>
    raw_message ${record["message"]}
  </record>
</filter>

<filter audit>
  @type parse_audit_log #fluent-plugin-filter-parse-audit-log
</filter>

But if I do the record_transformer before the parse_audit_log my new raw_message field is lost, and if I do it after, then there is no message field to copy from any more so it is set to null.

In an ideal world, I would have three fields under the json payload like:

jsonPayload: {
raw_message: "type=SYSCALL msg=audit(1611581847.881:273): ........."
body: {1}
header: {2}

Thanks