community
cancel
Showing results for 
Search instead for 
Did you mean: 

alteryx server Discussions

Find answers, ask questions, and share expertise about Alteryx Server.

scale out workers

Meteor

Dear Alteryx Community,

 

I am planning to install a full server on on VM (schedule, designer, worker, controller, mongoDB) and add a second VM with only a worker installation.

is that possible? If yes, is it a good way to scale out workers?

 

Any information will be appreciated.

Many thanks

Asma

Alteryx Certified Partner

Alteryx produced this YouTube video that has a good overview of scaling starting at 7:34.

https://www.youtube.com/watch?v=40BQTDWa5ec&t=434

 

Here's the official documentation:

http://downloads.alteryx.com/Documentation/Alteryx%20Server%20Installation%20Guide.pdf

 

The first thing that typically needs scaling is the workers. It sounds like you're good to get started, but your next move will probably be adding another worker to take the worker load from the server controller/db machine. 

Asteroid

Hi @asma_rez,

 

I think it is a good start to scale out of worker.

Actually, It depends on how many resource you have.

E.g.How many VMs you have and how many Core and RAM of each VM.

 

 

https://community.alteryx.com/t5/Engine-Works-Blog/Measuring-and-Scaling-a-Private-Server/ba-p/8786

 

Best regards,

Samuel

Alteryx
Alteryx

There isnt a single clear-cut answer to this question, because how and what you scale really depends on how the Server is being utilized.

 

That said, what you've described is in line with what I typically guide my customers to do - in almost all situations, Worker Bandwidth is the first constraint a server stack runs into and when you scale out for worker bandwidth, what you're describing is correct.

 

The attached is a diagram that I use often to explain the process.

 

I would also add that on VM's you will want to be careful with RAM pooling, as this can cause the appearance of CPU issues, when in fact you just dont have RAM available at the right time. Since most Alteryx Processing is done in-memory (in-db tools excluded), you want to make sure RAM is never a constraint.