A Big Data analytics company has deployed a stream processing application using KCL to read records from Kinesis Data Streams configured with multiple shards. The application is running on one EC2 instance. It seems that the consuming application is lagging under a large load and therefore records are not processed in time and eventually dropped from the stream. As a DevOps Engineer, you have been tasked with improving the reliability of this application with minimal changes, what should you do? (Select two)