Are there any views of the pros/cons of running a native NiFi cluster versus a 
cluster of standalone NiFi nodes (managed by puppet/chef etc to ensure they are 
configured consistently) fronted by software load balancer(s).  It is assumed 
the entry point to the graph will be an http listener.

Thoughts that spring to mind are it may be simpler to achieve canary and 
rolling deployments of graph/processor changes with a set of standalone NiFi 
instances behind a load balancer.

Many Thanks,

Nigel

CGI IT UK Limited.  A CGI Group Inc. Company
Registered Office: 250 Brook Drive, Green Park, Reading RG2 6UA, United 
Kingdom.  Registered in England & Wales - Number 947968

Reply via email to