Your operations team is facing a performance issue with a production application running on Google Compute Engine. Under heavy load, the application is dropping incoming requests. The process list for affected instances shows that a single application process is consuming all available CPU resources, and the autoscaling has already reached its maximum limit of instances. Other related systems, including the database, are not experiencing any abnormal load. To ensure that production traffic can be served again as quickly as possible, which action should you recommend?