Change settings in a controlled way

Question and objective

Which configuration affects the current result or behavior? The objective is to make changes in the right place and understand their effects.

Implementation in Bytefabrik

There is no single setting for every analysis. Connection parameters belong to the adapter, field mappings to processing, queries to the chart, and equipment rules to the relevant Insights configuration. Platform-wide administration and additional service configuration complement these resource-specific settings.

AI features also require the relevant services and model connections to be configured in the installation. Available options depend on the module and environment.

Best practices for changes

1. Identify the affected resource and its dependent analyses.
2. Record the original configuration and expected effect.
3. Test the change using a known data example.
4. Check behavior after saving or performing any required restart.
5. Inform affected users if the meaning of a metric changes.

What to consider

Changing a time filter is different from changing event time. A new target value can change a metric without improving the production process. Evaluate configuration changes and process changes separately.

For AI configurations, model connectivity and code execution are separate components. Check the actual setup rather than assuming where every service runs based on the platform's deployment location.

Controlled configuration makes analyses traceable and troubleshooting faster when values differ after a change.