who said anything about tracing? it sounds to me like mogens was recommending that you establish reasonable response time expectations for the sql in your database and then monitor that through the v$ views. possibly a trace to establish a baseline of resource consumption and to have a documented, acceptable execution path. ...when things go awry you have a (possible) plan to aspire to. ...but what if changes in your data warrant a change in execution? On 11/22/05, Juan Carlos Reyes Pacheco <juancarlosreyesp@xxxxxxxxx> wrote: > Mogens, > I understand the advice is to trace, this means gather statistics now > and trace now, and every month for example and see if execution plans > goes worst, if they goes worst then is time to gather statistics > again? > Thank you :) > -- > //www.freelists.org/webpage/oracle-l > > > -- //www.freelists.org/webpage/oracle-l