I’m new to Druid, and I’m trying to understand the autoscaling related configs, specifically “autoscale.workerVersion”:

If set, will only create nodes of set version during autoscaling. Overrides dynamic configuration.

I don’t understand what this config does, what the description means, or what to set it to if I am in fact running index nodes in an auto scaling group on EC2. Can someone elaborate on this?

Thanks in advance.


Hi Brian,

The config doesn’t have anything to do with autoscaling groups and is in fact kind of an alternative to autoscaling groups. It’s part of Druid’s built-in autoscaling code, which currently only applies to its middleManager nodes. The overlord has an ability to create and terminate them in response to changing load conditions.

Got it, thanks. Gian.