You are managing several internal applications that are deployed on Compute Engine. Business users inform you that an application has become very slow over the past few days. You want to find the underlying cause in order to solve the problem. What should you do first?

To diagnose a performance issue, the first step is to gather information. Inspecting logs in Cloud Logging and metrics in Cloud Monitoring provides crucial insights into the application's behavior, resource utilization (CPU, memory, disk I/O, network), and potential errors. This data helps identify bottlenecks or recent changes that might be causing the slowdown. Changing the machine type is a potential solution if resource exhaustion is identified, but it's premature without diagnosis. Restoring a database backup is a drastic measure for data corruption or loss, not a first step for performance issues, and might lead to data loss. Deploying on a Managed Instance Group with autoscaling and a load balancer is a scalability solution, not a diagnostic step for an existing performance problem, and would involve significant architectural changes.