Poor BI Performance hurts Portal Part - II

consider troubleshooting a Portal system downtime and you are unable to bring up the server ,you can log on to NWA and try to access the java system reports and notice the user activities or the slowest response time objects. you might find that one of the sever processes in yor instance , central instance or the dialog instance may be doing a garbage collection and using up all resources to bring up the server to function correctly.the system can be so slow that if you are running on a web dispatacher within the LAN, you could time out or the application could ( /irj ) time out due to a long delay in response from the server process. the next place to go is the visual administration tool which can tell you the costly object which is killing the server.on the portal, several systems that are connected to it may cause the system to come down. for example if a BI system is conneceted to the portal and if you run a scheduled report using BEx Broadcaster you can turn on a background job in BI and if the job fails for some reason you could end up with the background work process in BI running forever and this could potentially eat up resources in AS JAVA and hence bring the system down. hence, Portal need not be the one stop to identify bottlenecks. the interfaces play a huge role in its activities which goes back to the definition of Portal: an interface or one stop to access everything by using connectors to communicate with the interfaced systems.
this is the second part of the series. read the first one here