Capacity: Instances or Weighted

Each Elastigroup has a dedicated setting to control the group Capacity. The Capacity of the group provides the option to control the minimal, maximal and the target numerical value representing the number of instances in the group. The amount can be a direct number of instances or a weighted representation of vCPUs.

When you launch an Elastigroup, you can define the capacity units that each instance type would contribute to your application's performance and availability. By using Instance Weighting you can now define more flexible scaling rules and to maximize overall cost efficiency while putting zero DevOps efforts.

Normally, when an application scales, it typically does that in instance units, meaning, one or more Instances at a time. In most of the cases – it is the same instance type or size, for example, my application scales from 5 instances of c3.2xlarge to 15 instances of c3.2xlarge at peak hours.

When you use the Elastigroup weighting feature, you can define a target vCPU capacity for your application and specify a vCPU count.

How to enable Weighted Configuration

1. Create a new Elastigroup.
2. Complete the required information in the General and Compute configuration tabs.
3. In the Predictive Rebalancing tab, set Capacity Unit to vCPU.
4. Complete the rest of the configuration you need for the Elastigroup and click Create.

Elastigroup calculates the number of instances to launch by dividing the target capacity by the instance weight. If the result is not an integer, Elastigroup rounds it up to the next integer so that the size of your cluster is not below its target capacity. But first and foremost, Elastigroup will select the best server size or family according to their overall availability, pricing, and durability.

You can now run, scale, and manage heterogeneous clusters with ease.

tip

Using weighted vCPU capacity units is not supported for autoscaler enabled Elastigroups such as Kubernetes, ECS, Nomad.

Advanced Custom Weights

To allow this option, the Group Capacity should be set using the vCPU unit.

If you want instances to be weighted based on custom units – you can add the weights parameter to the Elastigroup configuration JSON. This will allow you to gain higher accuracy in selecting the instances that will operate the workload. This can be based on ECU or available Memory.

tip

If a running instance is not configured with a custom weight, its weight will be determined by the number of its vCPUs.

Elastigroup configuration example:

{
  "group": {
    "compute": {
      "instanceTypes": {
        "weights": [
          {
            "instanceType": "m4.xlarge",
            "weightedCapacity": 16
          },
          {
            "instanceType": "m4.2xlarge",
            "weightedCapacity": 32
          },
          {
            "instanceType": "m4.4xlarge",
            "weightedCapacity": 64
          },
          {
            "instanceType": "m4.10xlarge",
            "weightedCapacity": 160
          },
          {
            "instanceType": "m3.xlarge",
            "weightedCapacity": 15
          },
          {
            "instanceType": "m3.2xlarge",
            "weightedCapacity": 30
          },
          {
            "instanceType": "c4.2xlarge",
            "weightedCapacity": 15
          },
          {
            "instanceType": "c4.4xlarge",
            "weightedCapacity": 30
          }
        ]
      }
    }
  }
}