On-Premise SDLC Best Practices

This article defines recommended best practices for:

• Promoting the Odin AI Platform platform (application) from Non-Production to Production
• Promoting Odin AI Platform projects and AI use cases from Non-Production to Production
• Clearly defining roles and responsibilities for each promotion activity, including updates, testing, configuration, and approval

The approach is designed for On-Premise environments and aligns with common SDLC, security, and operational governance standards. This approach:

• Clearly defines roles and responsibilities during promotion
• Accounts for environment-specific configuration updates
• Aligns with enterprise On-Premise SDLC and security models
• Maintains strict Production access control
• Enables repeatable and auditable promotions

---

​

Environment Model (On-Premise)

Your team maintains separate On-Premise Non-Production and Production environments for Odin AI Platform. These environments are:

• Logically and physically isolated
• Promoted via controlled, manual processes

Both platform updates and project promotions are governed under a single SDLC framework.

---

​

SDLC Phases and Promotion Overview

​

Phase 1 – Design & Build (Non-Prod)

• Use cases are designed and configured in Odin AI Platform Non-Prod
• Prompts, workflows, and integrations are iterated and tuned
• Test data is used where possible

Owner: Business Team / SMEs
Support: Odin AI Platform (AA)

​

Phase 2 – Testing & Validation (Non-Prod)

• Functional testing and UAT
• Accuracy, performance, and security validation
• Business sign-off confirming readiness for promotion

Owner: Business Team / SMEs

​

Phase 3 – Promotion Readiness

Before any promotion:

• Change approvals are obtained
• Firewall and network requirements are reviewed
• Production access and RBAC approach is confirmed
• Environment-specific configuration values are identified

---

​

Promotion Mechanisms

​

Platform (Environment) Promotion

Platform promotion includes:

• Odin AI Platform version upgrades
• Security patches
• Platform-level configuration changes

​

High-Level Flow

1. Update applied in Non-Prod
2. Platform testing and validation
3. Change approval and scheduling
4. Deployment to Production
5. Post-deployment validation

​

Project / Use Case Promotion

Odin AI Platform supports promotion via a project export/import mechanism between On-Premise environments.

​

Included in Promotion

• Project configuration
• Workflows and agents
• Prompts and rules

​

Excluded from Promotion

• Users and role assignments
• Environment-specific secrets
• Network and firewall settings

This ensures Production access remains explicitly controlled.

​

High-Level Flow

1. Project finalized in Non-Prod
2. Export from Non-Prod
3. Import into Production
4. Environment-specific configuration updates
5. Production configuration and access setup
6. Smoke testing and go-live

---

​

Roles and Responsibilities for Promotion Activities

This section defines who performs each activity involved in promoting updates and configurations from Non-Prod to Prod.

​

Platform (Environment) Promotion Responsibilities

​

Key Point

Odin AI Platform (AA) is responsible for executing platform updates and running platform-level tests.
Run / Middleware teams control approvals, scheduling, and infrastructure readiness.

​

Project / Use Case Promotion Responsibilities

​

Key Point

Your team’s business stakeholders own validation and approval, while Odin AI Platform (AA) executes the technical promotion steps.

​

Post-Promotion Configuration (Production)

These activities occur after import to maintain Production controls.

---

​

Environment-Specific Configuration Updates (Required During Promotion)

When a project is promoted from Non-Production to Production, certain configuration values must be updated to reflect the Production environment. These values are not carried over automatically and must be reviewed during promotion.

​

RPA Configuration Updates (If Applicable)

For use cases that integrate with RPA tooling, the following values typically need to be updated in Production:

• Control Room URL
• Username
• User v2 Token

These values must reference Production RPA infrastructure and credentials.

​

Odin AI Platform Configuration Updates

Within Odin AI Platform, the following environment-specific values must be reviewed and updated after promotion:

• Base URL
  • Example:
    • Non-Prod: https://nonprod.example.com
    • Prod: https://prod.example.com
• Project ID
• Project xApiKey
• Project xApiSecret
• Smart Table ID(s)

These values are unique per environment and are required for correct runtime behavior in Production.

​

Responsibility for Configuration Updates

---

​

Production Operating Model

​

Your Team

• Owns Production operations
• Manages users, RBAC, and monitoring
• Provides first-line incident response

​

Odin AI Platform (AA)

• Owns platform stability and fixes
• Delivers platform upgrades and patches
• Supports complex configuration issues

Odin AI Platform does not act as the Production run team.

---

​

Change Management and Governance

• All Production changes follow formal change management processes
• Non-Prod is the only environment for experimentation
• Emergency fixes follow expedited approvals

Each Production use case should have:

• Business owner
• Defined scope
• Support contacts
• Rollback plan