Skip to main content

System Requirements

This page outlines the requirements for deploying and using Datalinx AI.

Deployment Options​

Datalinx AI can be deployed in several ways:

OptionDescriptionBest For
Snowflake Native AppRuns entirely within SnowflakeSnowflake-centric organizations
Cloud HostedManaged by Datalinx AIQuick start, minimal ops
Self-HostedDeploy in your infrastructureMaximum control, compliance

Snowflake Native App Requirements​

For the Snowflake Native App deployment:

Snowflake Account​

  • Snowflake account with ACCOUNTADMIN role access
  • Enterprise edition or higher (for Native Apps)
  • Compute warehouse with at least MEDIUM size

Required Privileges​

-- Grant required privileges
GRANT CREATE DATABASE ON ACCOUNT TO ROLE datalinx_admin;
GRANT CREATE WAREHOUSE ON ACCOUNT TO ROLE datalinx_admin;
GRANT CREATE COMPUTE POOL ON ACCOUNT TO ROLE datalinx_admin;

Network Requirements​

  • Outbound HTTPS access for:
    • External data sources
    • LLM providers (OpenAI, Anthropic) if using AI features
    • Reverse ETL destinations

Self-Hosted Requirements​

For self-hosted deployments:

Infrastructure​

ComponentMinimumRecommended
CPU4 cores8+ cores
Memory8 GB16+ GB
Storage50 GB SSD100+ GB SSD
Network100 Mbps1 Gbps

Software Dependencies​

SoftwareVersionPurpose
Docker24.0+Container runtime
PostgreSQL14+System database
Node.js18+Frontend build (dev only)
Python3.11+Backend runtime

Database Requirements​

Datalinx AI requires a PostgreSQL database for system metadata:

-- Minimum configuration
CREATE DATABASE datalinx;
CREATE USER datalinx_app WITH PASSWORD 'secure_password';
GRANT ALL PRIVILEGES ON DATABASE datalinx TO datalinx_app;

Recommended PostgreSQL configuration:

  • max_connections: 100+
  • shared_buffers: 25% of RAM
  • effective_cache_size: 75% of RAM

Cloud Provider Support​

Self-hosted Datalinx AI runs on:

  • AWS: ECS, EKS, or EC2
  • Azure: AKS or Azure Container Instances
  • GCP: GKE or Cloud Run
  • On-Premises: Any Docker-compatible environment

Network Requirements​

Inbound Ports​

PortProtocolPurpose
443HTTPSWeb UI and API
80HTTPRedirect to HTTPS

Outbound Access​

Datalinx AI needs to reach:

DestinationPurpose
Data sourcesConnect to your databases/APIs
LLM providersAI-powered features
Reverse ETL targetsPush data to destinations

Firewall Rules​

Allow outbound HTTPS (443) to:

  • api.anthropic.com (AI features)
  • api.openai.com (AI features)
  • Your data source endpoints
  • Your reverse ETL destinations

Browser Requirements​

Datalinx AI web interface supports:

BrowserMinimum Version
Chrome90+
Firefox90+
Safari14+
Edge90+
note

JavaScript must be enabled. Datalinx AI uses React for the frontend.

Data Source Requirements​

Database Sources​

Each database type has specific requirements:

PostgreSQL​

  • Version 12+
  • Read access to source schemas
  • pg_stat_statements extension (recommended for monitoring)

Snowflake​

  • Any current version
  • Warehouse with appropriate size
  • Read access to source schemas

Databricks​

  • Unity Catalog enabled (recommended)
  • Compute cluster or SQL warehouse
  • Read access to catalogs/schemas

API Sources​

For API data sources:

  • Valid OpenAPI 3.0 specification
  • Authentication credentials (API key, OAuth, etc.)
  • Stable endpoint URLs

File Sources​

For cloud storage sources:

  • Read access to buckets/containers
  • Properly formatted files (CSV, JSON, Parquet)
  • Appropriate IAM roles/service accounts

Security Requirements​

Authentication​

Datalinx AI supports:

  • Email/password authentication
  • SSO integration (SAML, OIDC)
  • Multi-factor authentication (MFA)

Encryption​

  • At Rest: All credentials encrypted using AES-256
  • In Transit: TLS 1.2+ required for all connections

Compliance​

Datalinx AI is designed to support:

  • SOC 2 Type II
  • GDPR
  • HIPAA (with appropriate configuration)
  • CCPA

Resource Sizing Guidelines​

Small Deployment (< 10 GB data)​

  • 2 CPU cores
  • 4 GB RAM
  • Small warehouse (Snowflake)

Medium Deployment (10-100 GB data)​

  • 4 CPU cores
  • 8 GB RAM
  • Medium warehouse (Snowflake)

Large Deployment (100+ GB data)​

  • 8+ CPU cores
  • 16+ GB RAM
  • Large warehouse (Snowflake)
  • Consider dedicated compute pools

Monitoring and Observability​

Datalinx AI exports metrics compatible with:

  • Prometheus
  • CloudWatch
  • Datadog
  • Application logs in JSON format

Next Steps​

Last updated on