Given an increase in latency corresponding with higher request volumes on a serverless compute service function, what would be the initial step to pinpoint potential improvements?
Activate tracing services to collect detailed execution data of the function.
Deploy a caching solution to offset the load on the compute resource.
Increase the memory size directly to the serverless function to handle more requests.
Begin decoupling processing steps via a fully managed message queuing service.