Workers

[Russ Wilson <rp...@gmail.com>]

In docs i do see much information regarding configuration properties that
are in the config.py. I see there is one called "SUPERSET_WORKERS" will,
does this allow us to scale up to multiple nodes? or how do you use it?

Thanks
Russ

Re: Workers

[Maxime Beauchemin <ma...@gmail.com>]

This defines how many worker processes should be started when firing the
`superset runserver` command, same as `superset runserver -w 16` where 16
means 16 subprocesses serving pages. So it's the parallelism parameter on
that one node. It should probably roughly match the number of CPUs on that
node. Note that it's also possible to run Gunicorn on async mode to get
more concurrency.

You can fire up as many individual web servers as you need, as well as as
many celery workers as you need.

How you do that is probably defined by how you do that kind of service
deployment internally already (Chef/Puppet/Ansible/Kubernetes, ...).

Max

On Sat, Sep 2, 2017 at 1:20 PM, Russ Wilson <rp...@gmail.com> wrote:

> In docs i do see much information regarding configuration properties that
> are in the config.py. I see there is one called "SUPERSET_WORKERS" will,
> does this allow us to scale up to multiple nodes? or how do you use it?
>
> Thanks
> Russ
>