Effective change management policies are critical for ensuring smooth transitions between testing and product environments, minimizing disruptions, and maintaining the integrity of your data and systems. This article describes best content migration practices tailored to Metric Insights environments, as well as environment governance models and migration dependencies.
1. Dealing with Proof of Concept (POC) Environment
Firstly, you have to decide how to deal with the Proof of Concept environment. Determine if the POC environment will transition to a development or production environment or be scrapped to start fresh.
Factors to consider:
- Assess the relevance and quality of content created during the POC phase. Define if the content created in this phase is actually useful for the Users
- Evaluate the alignment of the POC setup with long-term governance policies. Check if the access permissions, structure of content, and metadata are correct
- Ensure Data Source configurations and naming conventions comply with organizational standards of transitioning the POC environment. Check to what production tool or database the POC environment is connected
2. General Content Migration Recommendations
Here are some overall recommendations that will lessen the probability of losing quickly if needed. Changes to custom code (e.g., Portal Pages, Email Templates), and security model access are done in non-production
Advantages:
- Ensures vetted use cases before production implementation
- Allows for flexible ongoing maintenance in production
Risks:
- Moderate governance complexity may slow changes
3.3. Restrictive Governance Model
This governance model ensures maximum control and compliance in regulated environments. Main features of this model are:
- Use cases are developed in a sandbox or development environment
- Thorough testing is performed in a UAT environment that mirrors production
- Only non-exportable changes are made directly in UAT
- Approved use cases are promoted to production
- No changes are made directly in production
- Upgrades are sequentially tested from lower environments to production
- One lower environment must mirror production but allow controlled changes
Advantages:
- Rigorous testing minimizes production risks
- High compliance for regulated industries
Risks:
- Potential delays in implementation due to stringent processes
4. Migration Dependencies
Every Object in the Metric Insights has a set of connections to other objects within the system. When it is migrated, it is migrated with all those connections. Those connections can be described through two types of objects:
- Foreign Entities: Objects referenced by the migrated object
- Reference Entities: Objects referencing the migrated object
During migration, ensure connected objects are properly handled:
- Choose appropriate migration options (Folder relationships, Portal Page variables, Custom Fields)
- Maintain naming consistency for Datasets, Categories, Folders, and Dimensions across environments
NOTE: Ensure the destination environment can accommodate the object structure and dependencies without breaking functionality.
4.1. Migration Restriction: Dimensions
The migration of the Dimension is complicated and has to be done with caution. If the Dimension is a foreign or reference object, its migration is restricted to prevent problems. To avoid complications, make sure to maintain consistency through environments for the following Dimension features:
- The Dimension Key Value type (e.g., “text” / “integer”)
- The Dimension type (e.g., Compound / Simple)
- The Parent, Primary, and Secondary Dimension
- The Include "Total" Dimension Value feature status
4.2. Data Source Consistency
All environments (development, UAT, production) should ideally point to the same source image/data platform. It is not necessary to point lower environments of Metric Insights to lower environments of Tableau, Power BI, etc.
Key recommendations:
- Ensure Data Sources across environments (e.g., Tableau, Power BI) share consistent naming conventions for seamless migration. If prod and non-prod environments both point to the exact same tool (e.g. Tableau), they need to have the same Data Source name (e.g. "Tableau - Production")
- Periodically audit Data Source configurations to ensure alignment