Time to Spin Up Additional Master or Worker Nodes

Is there an average amount of time to spin up additional master or worker nodes?

I understand that the number of nodes can be adjusted via the “Platform Administration” menu, but how long does it take to scale up or down the infrastructure?

Is there a difference between each node type or is it agnostic for the most part?

Also, is there a way to programatically scale infrastructure up or down?

Context is to scale up additional infrastructure to support various use cases when requested and scale down when not needed to keep costs in check.

Thanks.

1 Like

Hey there,

  1. I’ve seen it take anywhere from 2minutes to 10 minutes to bring up a new node, depending on the specific facts of the environment.

  2. For worker nodes, a good strategy is to use C3’s build-in auto-scaling feature. If you are a cluster admin you can use the: Cluster.setAutoScale* family of methods to control this.

  3. For master nodes, there is no autoscaling. However, the Cluster.start* functions can be used to add masters.

Generally, custoemrs to not have the ability to call these methods since there is some danger of impacting either SLA (too few nodes) or cost (too many nodes) and these 2 metrics are quite sensitive :slight_smile: