What approaches are currently preferred for configuring incoming traffic: using the built-in ingress resource or explicitly specifying an ingress controller within the application?

In modern DevOps practices, it is usually preferred to explicitly specify an ingress controller outside the application rather than using an embedded ingress resource within the application.

Reasons:

• Separation of responsibilities: The ingress controller is managed at the infrastructure or cluster level, while the application remains independent of specific routing configurations.
• Flexibility and scalability: One ingress controller can serve multiple applications, simplifying management and updates.
• Security and centralized control: Security policies, SSL, routing, and load balancing are configured centrally.

Thus, a separate ingress controller (e.g., NGINX Ingress Controller) is usually created, and ingress resources are configured to reference application services, rather than embedding ingress logic inside the application.