Hari,

I echo John's question here.  I don't see any requirements on the wiki that 
require more than a syslog appender for log4j.  What this means is that 
whatever is logged to our current log file will get sent to syslog.  That's 
something someone can configure today on existing releases.  Do you have more 
use cases?  For example, is there anything that should be logged to syslogs but 
not in our logs or vice versa?

--Alex  

> -----Original Message-----
> From: John Kinsella [mailto:[email protected]]
> Sent: Wednesday, December 26, 2012 1:53 PM
> To: [email protected]
> Subject: Re: [DISCUSS] Syslog enhancements
> 
> (Changed subject as noted by Alex)
> 
> Question - is this feature something beyond using the syslog appender in
> log4j?
> 
> One thing I'd like to see is logs using key-vaue pairs. The closer to that we 
> can
> get, the easier it is for me to have the logs consumed by a separate analytics
> package.
> 
> One nitpick - syslog can be udp or tcp.
> 
> On Dec 26, 2012, at 11:12 AM, Hari Kannan <[email protected]> wrote:
> 
> > Hello All,
> >
> >
> >
> > I wish to propose syslog enhancements in CloudStack - I have added some
> details
> here<https://cwiki.apache.org/confluence/display/CLOUDSTACK/syslog+en
> hancements> along with a JIRA ticket 772
> >
> 

Reply via email to