Scale Edge Orchestrator#

You can deploy the Edge Orchestrator on cloud providers at various scales to support different number of edge nodes and you can scale Edge Orchestrator dynamically after the initial deployment; this allows for right-sizing of the Edge Orchestrator as the number of edge nodes grow, to achieve better cost efficiency.

Intel supports both horizontal and vertical scaling for general and observability nodes. You can achieve horizontal scaling by adding more Amazon Elastic Kubernetes Service* (Amazon EKS*) nodes to Edge Orchestrator, and vertical scaling by increasing the resources of the existing Amazon EKS nodes.

Intel also supports scaling the Aurora Capacity Units (ACU).

This section explains the available parameters and predefined scale profiles, and shows how to update cluster settings using the same provision script used for installation.

Scale Parameters#

The following table lists all available parameters that can be used to scale the cluster. These parameters are optional and can be supplied when calling the provision.sh update-cluster-setting script.

General#

Flag

Description

Default Value

Available Values

--profile {CLUSTER PROFILE TO DEPLOY}

Specify the cluster scale profile to deploy. This can be overridden if any of the above flags are provided.

default

default, 100en, 500en, 1ken
Compute#

Flag

Description

Default Value

Available Values

--desired-nodes {NUMBER OF NODES}

Specify the desired number of nodes for the cluster.

3

Any integer that is greater than or equal to the minimum number of nodes, and less than or equal to the maximum number of nodes.

--min-nodes {NUMBER OF NODES}

Specify the minimum nodes for the cluster.

3

Any integer that is less than or equal to the maximum number of nodes.

--max-nodes {NUMBER OF NODES}

Specify the maximum nodes for the cluster.

3

Any integer that is greater than or equal to the minimum number of nodes.

--node-type {Node type}

Specify the node type of the cluster.

t3.2xlarge

Any valid Amazon EC2 instance type.
Observability#

Flag

Description

Default Value

Available Values

--desired-o11y-nodes {NUMBER OF NODES}

Specify the desired number of observability nodes.

1

Any integer that is greater than or equal to the minimum number of observability nodes, and less than or equal to the maximum number of observability nodes.

--min-o11y-nodes {NUMBER OF NODES}

Specify the minimum number of observability nodes.

1

Any integer that is less than or equal to the maximum number of observability nodes.

--max-o11y-nodes {NUMBER OF NODES}

Specify the maximum number of observability nodes.

1

Any integer that is greater than or equal to the minimum number of observability nodes.

--o11y-node-type {Observability node type}

Specify the observability node type.

t3.2xlarge

Any valid Amazon Elastic Compute Cloud* (Amazon EC2*) instance type.
Database#

Flag

Description

Default Value

Available Values

--num-rds-instance {Number of RDS instances}

Specify the number of Amazon RDS instances.

1

Any integer.

--min-rds-acu {NUMBER OF ACU}

Specify the minimum number of ACU for Amazon RDS.

0.5

Any number.

--max-rds-acu {NUMBER OF ACU}

Specify the maximum number of ACU for Amazon Relational Database Service* (Amazon RDS*).

2

Any number.

Predefined Profiles#

You can also select from the following predefined profiles when provisioning the cluster:

Predefined Profiles#

Profile

Edge Nodes(up to)

Number of Compute Nodes

Compute Node Type

Number of Observability Nodes

Observability Node Type

Min/Max Units of Amazon RDS’s ACU

RDS Instances

Default Volume Size for Each Node (GB)

default

50

3

t3.2xlarge

1

t3.2xlarge

0.5/2

1

20

100en

100

3

t3.2xlarge

1

r5.2xlarge

0.5/2

1

128

500en

500

3

t3.2xlarge

1

r5.8xlarge

0.5/4

1

128

1ken

1,000

3

m4.4xlarge

1

r5.12xlarge

0.5/8

1

128

See Open Edge Platform System Requirements for more information about resource requirements for each scenario.

Update Cluster Scale Settings#

Note

Scaling Edge Orchestrator cluster may require changing of observability profile. For details on observability profiles, see Observability Configuration.

Note

When updating the cluster scale, you also need to provide flags from the installation step; you can find the required flags from the section on Installation Options

There are three ways you can specify scale settings:

  1. Use a predefined profile:

    ./provision.sh update-cluster-setting --profile 100en \
    --aws-account 1234567890 --customer-state-prefix customer1-bucket
    --environment customer1-env1 --parent-domain [root-domain].com --region us-east-1
    --email myname@abc.[root-domain] --socks-proxy proxy-dmz.example_proxy.com:1080

  2. Use flags to change the scale:

    ./provision.sh update-cluster-setting --desired-nodes 3 --desired-o11y-nodes 1 \
               --min-nodes 3 --max-nodes 3 \
               --min-o11y-nodes 1 --max-o11y-nodes 1 \
               --min-rds-acu 0.5 --max-rds-acu 2 \
               --num-rds-instance 1 \
               --node-type t3.2xlarge --o11y-node-type m4.4xlarge \
    --aws-account 1234567890 --customer-state-prefix customer1-bucket
    --environment customer1-env1 --parent-domain [root-domain].com --region us-east-1
    --email myname@abc.[root-domain] --socks-proxy proxy-dmz.example_proxy.com:1080

  3. Override certain values from the profile, for example, to have more nodes with the 100en profile:

    ./provision.sh update-cluster-setting --profile 100en --max-nodes 5 --desired-nodes 5 \
    --aws-account 1234567890 --customer-state-prefix customer1-bucket
    --environment customer1-env1 --parent-domain [root-domain].com --region us-east-1
    --email myname@abc.[root-domain] --socks-proxy proxy-dmz.example_proxy.com:1080