Preparing your AWS account

Prerequisites

Before preparing your cloud account, make sure that you're assigned the following AWS managed policies or an equivalent custom policy granting full access to resources:

• arn:aws:iam::aws:policy/IAMFullAccess

• arn:aws:iam::aws:policy/ServiceQuotasFullAccess

EDB Postgres AI Cloud Service requires you to check the readiness of your AWS account before you deploy your clusters. The checks that you perform ensure that your AWS account is prepared to meet your clusters' requirements and resource limits, such as:

• Is the AWS CLI configured to access your AWS account?
• Is there a sufficient limit on the number of vCPUs and Network Load Balancers (NLBs) left in your region?

Check AWS resource limits for running Cloud Service

EDB provides a shell script, called biganimal-csp-preflight, which checks whether requirements and resource limits are met in your AWS account based on the clusters you plan to deploy.

1. Open the AWS Cloud Shell in your browser.

2. From the AWS Cloud Shell, run the following command:

   → WrapCopy

   curl -sL https://raw.githubusercontent.com/EnterpriseDB/cloud-utilities/main/aws/biganimal-csp-preflight | bash -s <AWS-account-ID> <region> [options]

   The required arguments are:

   Argument	Description
   <account-id>	AWS account ID of your Cloud Service deployment.
   <region>	AWS region where your clusters are being deployed. See Supported regions for a list of possible regions.

   Possible options are:

   Options	Description
   -h or --help	Displays the command help.
   -i or --instance-type	AWS instance type for the Cloud Service cluster. The help command provides a list of possible VM instance types. Choose the instance type that best suits your application and workload. Choose an instance type in the memory optimized R5, R5B, or R6I series for large data sets. Choose from the compute-optimized C5 or C6I series for compute-bound applications. Choose from the general purpose M5 or M6I series if you don't require memory or compute optimization.
   -a or --high-availability	DEPRECATED - Enables high availability for the cluster. See [Supported cluster types(../../overview/02_high_availability) for more information.
   -x or --cluster-architecture	Defines the Cluster architecture and can be single, ha, or eha. See Supported cluster types for more information.
   -n or --networking	Type of network endpoint for the Cloud Service cluster, either public or private. See Cluster networking architecture for more information.
   -r or --activate-region	Specifies region activation if no clusters currently exist in the region.
   --onboard	Checks if the user and subscription are correctly configured.

   The behavior of the script defaults to --onboard if you provide no other options.

   For example, if you want to deploy a cluster in an AWS account having an ID of 1234-5678-9012, with an instance type of r5.24xlarge, in the us-east-1 region, in a public endpoint, and with no existing cluster deployed, run the following command:

   → WrapCopy

   curl -sL https://raw.githubusercontent.com/EnterpriseDB/cloud-utilities/main/aws/biganimal-csp-preflight | bash -s \
   1234-5678-9012 \
   us-east-1 \
   --instance-type r5.24xlarge \
   --networking public \
   --activate-region \
   --onboard\

The script displays the following output:

• Whether your AWS account restricts vCPUs, elastic IP addresses, VPCs, or NLBs in your region (and availability zone, if HA is enabled). Open an AWS support request to remove restrictions for the resources with NotAvailable displayed in the Suggestion column. See Request quota increase. For default service quota limits in AWS, see AWS service information

  → WrapCopy

  ##############################################
  Checking Service Quotas Limits on us-east-1...
  ##############################################
  
  Resource              Quota Name                            Limit    Used     Required    Gap      Suggestion
  --------              ----------                            ----     -----    --------    ---      ----------
  m5(a).large vCPUs     Running On-Demand Standard instances  512      0        6           410      OK
  r5.24xlarge vCPUs     Running On-Demand Standard instances  512      0        96          410      OK
  Elastic IP Addresses  EC2-VPC Elastic IPs                   5        0        3           2        OK
  VPCs                  VPCs per Region                       5        1        2           2        OK
  NLBs                  Network Load Balancers per Region     50       0        1           49       OK
  
  Note: the first two Instance Types are referring to the same AWS Service Quota.

Configure your AWS account

Open an AWS support request to remove restrictions for the resources with NotAvailable displayed in the Suggestion column. See Request quota increase. For default service quota limits in AWS, see AWS service information.