Hi Santosh,
Basically all RFC processes are run in a dialog work process and thus possibly creating shortages of dialog work processes for normal dialog users. In general though the system should be set up, that the RFC processes get only a limited amount of resources. There are quite a few profile parameters for that. On top of that it's also application design, as it matters how the RFC calls itself are issued (e.g. on which server are they executed, etc.).
In some cases the automated restart of failed tRFC's can wreak major havoc on your job schedule. I.e. if you have tons of failed tRFC calls queued up in the outbound RFC queue and the job scheduler has to try to restart all of them, your normal jobs will get affected (as job scheduler checks usually only every minute for outstanding jobs and if you have tons of short running jobs it'll take long time before they disappear).
I don't think anybody can give you a general hint how to resolve RFC performance issues. First of all, performance as always is relative and means prioritizing the usage of system resources. Depending on your specific requirements, the way the applications work and what issues you encounter you will have to implement appropriate measures. Your basis folks should be able to do a proper performance analysis and provide recommendations on how to proceed.
Cheers, harald
Bluesky
