AWS EKS Deployment Prerequisites
Prepare the following infrastructure separately for Staging and Production before deploying the AI Command Center. Review the deployment overview for scope and responsibilities.
Application capacity
The deployment baseline follows the Medium Deployment category in the AI Command Center system requirements. Staging uses the minimum configuration and Production uses the recommended configuration.
| Resource | Staging minimum | Production recommended |
|---|---|---|
| CPU | 8 cores, 2.5 GHz or faster | 12 cores, 2.5 GHz or faster |
| RAM | 16 GB | 24 GB |
| Application disk | 100 GB | 250 GB |
| Network | 1 Gbps | 1 Gbps |
| Container orchestration | Amazon EKS | Amazon EKS |
| Container registry | Amazon ECR | Amazon ECR |
These values describe application capacity, rather than individual worker-node specifications or Helm resource settings. Account for Kubernetes overhead and supporting workloads when selecting nodes. Database and optional ingestion infrastructure are provisioned separately. Adjust Production capacity for workload volume, concurrent executions, voice conversations, and availability requirements.
Kubernetes and AWS infrastructure
Provide operational EKS environments with sufficient compute capacity and deployment permissions, customer-managed ECR repositories, internal service connectivity, an approved ingress or load balancer, DNS, TLS certificates, and writable persistent storage through a suitable StorageClass.
Databases
The Customer Infrastructure team provisions, secures, backs up, and operates the databases. Use separate databases and credentials for Staging and Production, with network access from the applicable application services.
| Database | Technology | Requirement |
|---|---|---|
| Primary platform database | PostgreSQL | Mandatory; stores platform data and execution history |
| Execution history archival database | PostgreSQL | Required when archival is enabled |
| Ingestion database | PostgreSQL | Required when Ingestion is enabled |
| Vector database | Qdrant v1.17.0 or above | Required when Ingestion is enabled |
| Graph database | Neo4j Server 5.26.x or above | Required when Ingestion is enabled |
Execution History Archival is recommended to control primary database growth. Depending on configured policies, eligible records may be moved to a separate archival database or permanently deleted. Provide a separate archival database in each environment when archival is enabled.
Provision ingestion dependencies externally and maintain independent configurations and data stores for each environment that uses Ingestion. Confirm release compatibility with IntelliBuddies before provisioning dependency versions.
Persistent configuration and storage
The application modifies appsettings.json during runtime. Store it on writable persistent storage so changes survive container restarts, Pod recreation, and upgrades. Provision initial configuration when the volume is first created, and protect sensitive values through access controls and secrets management.
Storage must support the application's Pod placement and shared access requirements. Use separate persistent storage for Staging and Production, with the required application read/write permissions. Amazon EFS may be considered for shared access across nodes, while Amazon EBS may suit single-writer storage, subject to Engineering validation.
Connectivity
Provide DNS resolution and any necessary firewall or proxy configuration for outbound HTTPS on TCP port 443.
| Endpoint | Purpose | Requirement |
|---|---|---|
https://accounts.intellibuddies.com | License activation | Mandatory |
https://registry.ib-x.ai | Download images for import into ECR | Required from the image import and update process |
https://igservices.intellibuddies.com | Integration Gateway services | Required when corresponding services are used |
https://supportanalytics.intellibuddies.com | Integration Gateway usage analytics | Required when corresponding services are used |
Application Pods use images from customer ECR and do not require direct access to the IntelliBuddies registry for image pulls. Additional connectivity depends on the configured agents, AI providers, and third-party integrations.
HTTPS and licensing
Secure external access with HTTPS. The customer configures certificates and TLS termination through the approved load balancer or ingress. Application containers expose port 8080 internally; external HTTPS access uses port 443. End-to-end TLS inside Kubernetes requires additional configuration and compatibility validation.
Activate a valid license for each environment before making it available for use. IntelliBuddies provides activation guidance; ongoing validation and renewal connectivity follow the applicable licensing configuration. See License Activation for general licensing information.