I'm using chainsaw 2.1.0 snapshot from http://people.apache.org/~sdeboy/
I was able to set it up to parse a log file, but it always splits multiline messages into multiple one-line messages (including all the other fields into the first message)
How can I make it parse multiline messages correctly? Is there any workaround or alternative?
Edit: a few lines from the file:
[<proceed>(2)] INFO 2013-01-16 05:04:23,728 xmlstream.py:start_tls:740 ::: Negotiating TLS :::
[<proceed>(2)] INFO 2013-01-16 05:04:23,729 xmlstream.py:start_tls:741 ::: Using SSL version: 3 :::
[<proceed>(2)] DEBUG 2013-01-16 05:04:23,838 xmlstream.py:start_tls:809 ::: CERT: -----BEGIN CERTIFICATE-----
MIIDaDCCAtGgAwIBAgIKMfrHUQAAAABbMDANBgkqhkiG9w0BAQUFADBGMQswCQYD
VQQGEwJVUzETMBEGA1UEChMKR29vZ2xlIEluYzEiMCAGA1UEAxMZR29vZ2xlIElu
dGVybmV0IEF1dGhvcml0eTAeFw0xMjA2MDUwMDUyMTRaFw0xMzA2MDUwMTAyMTRa
MGMxCzAJBgNVBAYTAlVTMRMwEQYDVQQIEwpDYWxpZm9ybmlhMRYwFAYDVQQHEw1N
...
Zfhf9Lio+EplEmBc
-----END CERTIFICATE-----
:::
and the log format is "[NDC] LEVEL TIMESTAMP FILE:METHOD:LINE ::: MESSAGE :::"
(I attempted to put unique separators before and after the message to help with parsing)
Try setting 'appendNonMatches' to false in the receiver configuration.
If that doesn't work, provide a few lines of your log file (a mix of single and multi-lines) and I can tell you what you need to do configuration-wise.
It might be a bug too, will have to see.
A few notes after your responses: