What is required before an API implemented using the components of Anypoint Platform can be managed and governed (by applying API policies) on Anypoint Platform?

In one of the critical payment related mule application, transaction is being used . As an enhancement to implementation , scatter gather route is introduced which is also the part of transaction group. Scatter gather route has 4 routes.
What will be the behavior of the Mule application in case of error occurs in 4th route of the scatter-gather router and transaction needs to be rolled back?

An API has been unit tested and is ready for integration testing. The API is governed by a Client ID Enforcement policy in all environments.
What must the testing team do before they can start integration testing the API in the Staging environment?

A mule application must periodically process a large dataset which varies from 6 GB lo 8 GB from a back-end database and write transform data lo an FTPS server using a properly configured bad job scope.
The performance requirements of an application are approved to run in the cloud hub 0.2 vCore with 8 GB storage capacity and currency requirements are met.
How can the high rate of records be effectively managed in this application?

An organization is designing a Mule application to periodically poll an SFTP location for new files containing sales order records and then process those sales orders. Each sales order must be processed exactly once.
To support this requirement, the Mule application must identify and filter duplicate sales orders on the basis of a unique ID contained in each sales order record and then only send the new sales orders to the downstream system.
What is the most idiomatic (used for its intended purpose) Anypoint connector, validator, or scope that can be configured in the Mule application to filter duplicate sales orders on the basis of the unique ID field contained in each sales order record?