Licensing, billable AWS services, environment sizes, cost control, quotas, key rotation, patches, upgrades and backups for a deployment.
This page describes how a Personalisation Hub deployment is licensed, sized, maintained, patched and backed up over its life in your AWS account.
Licensing and cost model
Item
Description
Licence
Purchased through the AWS Marketplace listing, which carries the current pricing, billing dimensions and support entitlements
Licence fee
Covers access to the software only
AWS infrastructure
Billed separately by AWS; not included in the licence fee
Licence binding
Tied to the AWS account that provisions the stack; no licence keys or manual activation
Cancellation
Managed through AWS Marketplace
Billable AWS services
Actual cost depends on Region, instance sizing, retention periods and usage.
AWS service
Mandatory
Purpose
Amazon EC2
Yes
Core application services and the Installation Wizard
Amazon EKS
Yes
Orchestrates containerised workloads
Amazon RDS (PostgreSQL)
Yes
Structured application data
Amazon S3
Yes
Assets, reports, logs and backups
Amazon ElastiCache (Redis)
Yes
Caching and session storage
Amazon MSK
Yes
Event and message queue backbone
Amazon OpenSearch Service
Yes
Log collection, indexing and analytics
AWS Certificate Manager
Yes
TLS/SSL certificates for HTTPS
AWS IAM
Yes
Roles, permissions and authentication
Amazon CloudWatch
Yes
Monitoring, alerting and log retention
AWS KMS
Yes
Encryption of sensitive data
Elastic Load Balancing (ALB/NLB)
Yes
Routes traffic to services
AWS NAT Gateway
Yes
Outbound internet access from private subnets
Amazon Route 53
Optional
Custom domain names and hosted zones
Environment size
The environment size is chosen in the Installation Wizard and can be changed later on its Cost Control page. It sets the EKS worker node instance type.
Environment size
EKS instance type
Intended for
Functional
t3.medium
Testing or proof of concept
Low
t3.large
Light production workloads
Medium
t3.xlarge
Typical production usage
High
t3a.2xlarge
High-traffic or compute-intensive use
If the instance type is unavailable in your Region, a comparable alternative is selected automatically. The Functional and Low sizes are not suitable for the AI features.
Cost control options
Setting
Where
Description
Instance types
Installation Wizard → Cost Control
Set by the environment size
Autoscaling thresholds
Installation Wizard → Cost Control
When nodes and pods scale
Log retention
Installation Wizard → Cost Control
How long logs are kept
Scale Down / Scale Up times
Installation Wizard → Cost Control
Daily shutdown and start-up times in the local time zone for non-critical environments; applies mainly to EC2-based services
Weekend scale-down
Installation Wizard → Cost Control
Optional; keeps the environment scaled down over weekends
Service quotas
Item
Description
Quotas that commonly apply
EC2 instances, EBS volumes, VPCs, KMS keys, RDS, EKS, ElastiCache, OpenSearch, NAT gateways and VPC components; defaults vary by Region and account
Check during installation
Automatic; if a quota is reached the stack pauses with a descriptive error
Raising a quota
Through the AWS Service Quotas console, ideally before installation
Resuming
In the Installation Wizard, select Try Again after the quota increase is granted
Credentials and keys
Item
Description
IAM users and access keys
Created by the stack (for example TerraformPermissionIAMUser) and used during installation; managed and rotated by you under AWS best practice
KMS customer-managed keys
Encrypt RDS, S3 and EBS; created by the stack, with automatic key rotation available in KMS
Manual creation
Not required; the stack creates and manages these resources
Fault handling
When
Behaviour
During installation
The Installation Wizard shows a detailed error. Once the cause (IAM permission, quota, networking) is fixed, Try Again resumes the installation
After deployment
Service status is visible in the AWS Console: EC2 instance health, EKS workloads, RDS and ElastiCache availability, OpenSearch domain status. Failed components can be rebooted or redeployed; unreachable endpoints point to VPC, DNS or IAM settings — see AWS deployment: testing and troubleshooting