You are implementing an Azure solution that uses Azure Cosmos DB and the latest Azure Cosmos DB SDK. You add a change feed processor to a new container instance. You attempt to read a batch of 100 documents. The process fails when reading one of the documents. The solution must monitor the progress of the change feed processor instance on the new container as the change feed is read. You must prevent the change feed processor from retrying the entire batch when one document cannot be read. You need to implement the change feed processor to read the documents. Which feature should you use to prevent change feed processor from retrying the entire batch when one document cannot be read ?