In container orchestration, what is the fundamental purpose of a deployment in terms of desired state and updates?

In container orchestration, what is the fundamental purpose of a deployment in terms of desired state and updates?

• A. Expose services to the public Internet
• B. Maintain the desired number of pods and perform controlled updates over time
• C. Collect metrics
• D. Manage persistent storage volumes

Answer: (B)

The essential idea is that a deployment declares the desired state for your app and how updates should happen, and the orchestration system works to keep that state true over time. By specifying the number of pod replicas and the pod template, the deployment defines what the running system should look like. The system then continuously reconciles actual state with this desired state: it creates or deletes pods to reach the target replica count, restarts failed ones, and scales up or down as the spec changes. For updates, it uses a rolling-update approach that gradually replaces old pods with new ones, aiming to keep the application available throughout the transition. You can control rollout behavior and rollback if problems occur. In contrast, exposing services to the internet, collecting metrics, and managing persistent storage involve other components, not the deployment. So the primary purpose is to maintain the desired number of pods and perform controlled updates over time.

The essential idea is that a deployment declares the desired state for your app and how updates should happen, and the orchestration system works to keep that state true over time. By specifying the number of pod replicas and the pod template, the deployment defines what the running system should look like. The system then continuously reconciles actual state with this desired state: it creates or deletes pods to reach the target replica count, restarts failed ones, and scales up or down as the spec changes. For updates, it uses a rolling-update approach that gradually replaces old pods with new ones, aiming to keep the application available throughout the transition. You can control rollout behavior and rollback if problems occur. In contrast, exposing services to the internet, collecting metrics, and managing persistent storage involve other components, not the deployment. So the primary purpose is to maintain the desired number of pods and perform controlled updates over time.