The use of NiFi has steadily grown over the last few months and we have more people creating data flows. An issue that we have is the quality of the data flows varies a lot. I was wondering if there were any best practise for creating data flows?
I'm guessing this is a issue lots of other people have faced.
This is a tough question to respond to meaningfully as it is pretty
open ended. You're right that often different flows created by
different people will varying levels of optimization, efficiency,
simplicity, and so on much like written code.
One of the aspects that helps here is the ability to visualize and
follow the logic of flows. Often, messy or complicated flows reveal
themselves quite obviously and this can also help. That said, the
community has recognized a need for a more effective centralized
configuration management mechanism and this is the impetus behind the
Apache NiFI Registry subproject. This is where things like flow
versions, extensions, and potentially other artifacts can be centrally
stored and versioned and accessed via a variety of nifi environments
which will also aid learning and building upon the work of others.
Today we offer templates but they simply cannot be as useful as this
registry concept will be.
> The use of NiFi has steadily grown over the last few months and we have more
> people creating data flows. An issue that we have is the quality of the data
> flows varies a lot. I was wondering if there were any best practise for
> creating data flows?
> I'm guessing this is a issue lots of other people have faced.
> I have found the following documentation:
> but I was wondering if there was anything else?