Apply suggestions from code review

Co-authored-by: Thomas Gran <[email protected]>

Author: leonardehrenfried

docs/Migrating-Configuration.md

@@ -16,22 +16,19 @@ to be a configuration error, perhaps because the schema was changed. In such a c
 consult the [feature](Configuration.md#otp-features), [router](RouterConfiguration.md) or 
 [build configuration documentation](BuildConfiguration.md) to find out what the new schema looks like.
 
-By default, OTP starts up even when unknown configuration parameters have been found. This is there
-to support the style deployment where old config parameters remain in the file for a certain migration 
-period.  
-
-This allows you to roll back the OTP version without the need to roll back the OTP configuration. 
+By default, OTP starts up even if unknown configuration parameters exists. This supports forward and backwards
+migration of config parameters independent if the OTP version. This allow the configuration to follow the lifecycle of the environment and still be able to roll back OTP. Combined with the config parameter substitution it also allows using the same configuration in different environments. Tip! Rolling out the configuration before rolling out a new
+version of OTP, ensure you that you are safe and can roll back later. 
 
 An example: you change the location of the graphs, a critical error occurs afterwards and you need to 
 roll back. Any member of the dev-ops team should then be confident that they can roll back without 
-risk - even if the environment changed.
+risk - even if the environment changed since the last OTP version was rolled out.
 
 ### Aborting on invalid configuration
 
 If you want OTP to abort the startup when encountering unknown configuration parameters, you can add 
 the flag `--abortOnUnknownConfig` to your regular OTP CLI commands.
 
-This should of course be used wisely as it can also accidentally make your production instances refuse 
-to start up.
-Therefore, we recommend that you use it only in a separate pre-production step, perhaps during graph
-build.
+This should of course be used wisely as it can also accidentally make your production instances refuse to start up.
+For some deployments this is a good solution - especially if the config substitution feature is used to inject
+environment specific information. Using this feature in the graph-build phase is less risky, than in the OTP serve phase.