You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@nifi.apache.org by Anurag Sharma <an...@gmail.com> on 2020/02/12 08:08:45 UTC

Machines for NiFi: general purpose or memory intensive

Hi All,

We use  Digital Ocean, just wish to check with the community if we should
go for general purpose machines or memory-intensive machines for our NiFi
cluster?

Regards
Anurag

Re: Machines for NiFi: general purpose or memory intensive

Posted by Mike Thomsen <mi...@gmail.com>.
Most of your pain points with performance will be either CPU or disk I/O
unless you happen to have a particularly memory-intensive flow (most
aren't). I'd start with a heap of about 4GB of RAM/node and scale up from
there as needed while testing. Odds are pretty good that you'll find that
adding cores, not a lot of memory, will be the source of most improvements.
Start smallish like that and figure out how performant the process must be.
My main client, for example, prefers to spend less on cloud resources and
accept the longer wait as a way of saving money since there's no business
rush to get the data in really fast.

On Wed, Feb 12, 2020 at 3:09 AM Anurag Sharma <an...@gmail.com>
wrote:

> Hi All,
>
> We use  Digital Ocean, just wish to check with the community if we should
> go for general purpose machines or memory-intensive machines for our NiFi
> cluster?
>
> Regards
> Anurag
>