It looks like filebeat didn't handle > character properly and displayed the Unicode hex equivalent i.e. \u003c142 in output. Other fields are displayed just fine.
Any recommendations?
Filebeat version: 6.3
Input: SYSLOG or TCP [Both input types are having this problem]
Destination: Kafka
Regarding the syslog parser error, that seems to be caused by the format of the date. 2018-06-19 02:13:38 gets rejected but something like 2018-06-19T02:13:38 would be fine. I'm not a RFC3164 expert, but from looking at the parser and spec that seems to be the cause of the parse failure.
That's a problem for us. Syslog event displays event priority inside <>. Is there any way I can disable HTML escaping when marshaling JSON with filebeat?
Also, Is it possible to define datetime format inside filebeat config such as %Y-%m-%d %H:%M:%S
There isn't yet. Please comment on github issue to help move the discussion along.
The syslog parser is not configurable. But perhaps it can be enhanced to handle more date formats. My only worry is that it then might not be RFC compliant. What does @pierhugues think?
I think the following would be the correct mapping.
190 is the priority, 2018-06-19 02:13:38 should be the date. Local7.Info hostname HOST-1111-PQWER-SWT1 The software or appliance. 2018-06-19T02:13:38.635322-07:00 HOST-1111-PQWER-SWT1 Acl: %ACL-6-IPACCESS: list Isolation-4E Vlan100 denied tcp 10.10.10.10(56393) should be the message
The message parts contain another date, this is not part of the original Syslog RFC, but it could be extracted with a custom ingest pipeline.
The current date format trips the parser but adding support for it should not be too hard.
@pierhugues Yep... we're receving logs from a centralized syslog server and not from actual appliance (CISCO NX-OS). Also, mapping suggested by you should work.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.