Good morning. I have established an initial series of workflows for my customers that run on a single NiFi instance. We are running NiFi 0.7.1, and are bound to that release by external organizational dependencies.
My workflows have been exceeding expectations so far. Keeping pace with data flows - in fact, untaxed at all by the data flow demands we have experienced so far. I expect that situation to change though as more and more legacy processes migrate to our NiFi workflow. As the next step in the maturity of our NiFi capability I would like to see us eliminate the single-threaded nature of our architecture. I'd also like to position us to better scale horizontally as workflow demands increase. For these reasons I need to get us from where we are today to a clustered architecture. I'll most likely start simple: a Node Cluster Manager to manage our cluster, and two nodes. I am reading to learn how to migrate from what I have now to a clustered configuration. Wanted to also reach out to our user community here and ask: have any folks identified any "best practices" or "watch out for these gotchas" lists you can offer to help me build a better clustered solution? Thanks very much. -Jim
