Actually if you enable the properties shown in gateway-log4j.properties and
restart you should get request/response sizes in a separate access log but
unfortunately not the authenticated/mapped user. Keep in mind that the request
length might be based on the request Content-Length header. The code involved
is in gateway-server/src/main/java/org/apache/hadoop/gateway/trace and
AccessHandler.java in particular.
#log4j.logger.org.apache.hadoop.gateway.access=TRACE,httpaccess
#log4j.additivity.org.apache.hadoop.gateway.access=false
#log4j.appender.httpaccess=org.apache.log4j.DailyRollingFileAppender
#log4j.appender.httpaccess.File=${app.log.dir}/${launcher.name}-http-access.log
#log4j.appender.httpaccess.DatePattern=.yyyy-MM-dd
#log4j.appender.httpaccess.layout=org.apache.log4j.PatternLayout
#log4j.appender.httpaccess.layout.ConversionPattern=%d{ISO8601}|%t|%m%n
> On Dec 14, 2016, at 11:15 PM, larry mccay <[email protected]
> <mailto:[email protected]>> wrote:
>
> Hello Rajesh -
>
> I don't think that we have any way to do this today.
> There may be extension points in our metrics or in our audit logging to
> capture payload sizes.
> I imagine that the interesting place to do this would be in the dispatch
> classes to capture what is being sent to and received from the backend
> services.
>
> This would require some investigation to determine whether we can gather the
> sizes and how best to capture them.
>
> I would suggest filing an improvement JIRA and continuing the discussion
> there.
>
> Thanks for your interest in Knox!
>
> --larry
>
> On Wed, Dec 14, 2016 at 1:12 PM, Rajesh Chandramohan <[email protected]
> <mailto:[email protected]>> wrote:
> Hi Knox-team,
>
> In knox can we capture the bytes transferred information along with file name
> etc. This would help us to track how much transaction each users are doing
> via knox , Any method or thoughts on this?
>
> We make use of audit logs to capture each user how many transaction doing
> per hr per day. But we like to capture the details of how much bytes per
> transaction users are doing which will help us for various cases to manage
> Hadoop and Knox better. Any thoughts on this ?
>
>
> Thanks
> \Rajesh
>
>
