Thanks Guido, I hope that with the rsyslog magic (see startmsg.regex="^[[:digit:]]{2}-[[:alpha:]]{3}-[[:digit:]]{4} [[:digit:]]{1,2}:[[:digit:]]{1,2}:[[:digit:]]{1,2}\\.[[:digit:]]{3}" for the reference), I will be able to group this messages, otherwise I'll jump into your solution.
Joan Missatge de Jäkel, Guido <g.jae...@dnb.de> del dia dc., 22 de maig 2019 a les 8:55: > Dear Joan, > > by use of the common scripts, the file catalina.out will contain the > console output (stdout/tderr) of the *JVM* process and -- if not configured > in another way -- of the applications. Therefore, you have to deal with the > features of output formatting of current JVMs. Or -- as me -- you might > pipe this file descriptors through a tiny script that will prepend a > timestamp in a format of your choice. This will have the advantage that it > also handle "console" output of Java application; despite of the fact that > using stdout/stderr instead of a java logging mechanism is a very bad style. > > If you don't need to process thousands of lines per second, a simple shell > script may do the job: > > while read line; do echo "`date -Ins` $line"; done > > To avoid "double-stamping", you may add an heuristic check of the incoming > line; please adjust the RegExpr to your neeeds > > while read line; do [[ ! "$line" =~ ^\d\d\d\d ]] && echo -n > "[`date -Ins`] "; echo $line; done > > > Guido >