Hi Ravi, Thanks for your answer.
> One way we are currently doing this is: Expose limited set templates (those will work with pre-defined sources and destinations) to users in a custom portal, ask user specific configurations in the UI (custom) based on template selection and integrate this UI with NiFi using REST. as template users have the possibility to modify template configuration, if I understand your answer correctly, it means the only way is: 1) define custom processors with limited choices (to restrict the user choice possibilities) => custom dev (processor) to do 2) modify template upload UI mechanism, to go, not through a file (template) search but through a remote portal offering template search for upload => custom dev (UI + rest server) to do AFAIU this point (2) is what I understand from your sentence "ask user specific configurations in the UI (custom) based on template selection and integrate this UI with NiFi using REST". Do I am correct? Thanks. Regards, Dominique 2018-02-12 18:42 GMT+01:00 Ravi Papisetti (rpapiset) <[email protected]>: > One way we are currently doing this is: Expose limited set templates > (those will work with pre-defined sources and destinations) to users in a > custom portal, ask user specific configurations in the UI (custom) based on > template selection and integrate this UI with NiFi using REST. > > > > Thanks, > > Ravi Papisetti > > > > *From: *Dominique De Vito <[email protected]> > *Reply-To: *"[email protected]" <[email protected]> > *Date: *Monday, 12 February 2018 at 11:18 AM > *To: *"[email protected]" <[email protected]> > *Subject: *Best Pratices for implementing a source catalog for Nifi > > > > Hi, > > I just wondering what are the Best Pratices for implementing a source > catalog for Nifi. > > I mean the following: > > > * Nifi is able to catch data from various sources, and that's just fine > (there are plenty of processors!) > > * In a more constrained world, I would like to limit the sources a Nifi > (workflow) developer is able to select (and then, to connect to). > > So far I have not seen a way to do so. > > The single way I see (currently) is the following one: > > * derivate existing processors > > * when a (to be modified) processor offers a free-form choice, transform > it (if needed) into a choice into a list of N options (let's say for > example N possible Kafka clusters as sources, and no more than the > referenced ones). > > > > Is there another way than derivating the existing processors? > > > > Thanks. > > Regards, > Dominique > > > > >
