Cloud Digital Leader Google Cloud Products and Services Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Google Kubernetes Engine (GKE)

  Why it's wrong here

  GKE is a managed Kubernetes service that runs on pools of Compute Engine VMs. Even with cluster autoscaling enabled, a GKE cluster cannot scale to zero nodes because system pods (kube-system, networking, DNS, etc.) must run on at least one node at all times. This means you pay for node infrastructure even during idle periods, and you are responsible for node pool maintenance, upgrades, and capacity management. For a lightweight microservice that only needs occasional invocations, GKE is overkill compared to a fully serverless option.

• ✓

  Cloud Run

  Why this is correct

  Cloud Run is a serverless compute platform that executes stateless HTTP-driven containers on demand. When no requests are in flight, the service scales down to zero instances, so you are not billed for idle resources. It automatically provisions and scales instances up to handle spikes, supports any language and arbitrary Docker images, and only charges for the exact number of requests processed, making it a natural fit for a containerized microservice with intermittent traffic.

• ✗

  Compute Engine with managed instance groups

  Why it's wrong here

  Compute Engine managed instance groups can autoscale based on CPU utilization or HTTP load, but the group maintains a minimum instance count—by default at least one—so the VM stays running 24/7. To scale to zero, you would need to manually intervene or set up a custom monitoring/automation solution, which is not a native autoscaling behavior. Each VM also needs periodic OS patching and ongoing cost even at idle, making this option less suitable than a fully managed serverless service like Cloud Run.

• ✗

  App Engine Standard environment

  Why it's wrong here

  App Engine Standard is a serverless platform, but it only supports a fixed set of predefined runtime sandboxes, such as Python, Java, Node.js, and Go. It does not permit you to supply arbitrary Docker containers; instead, applications must conform to the platform's runtime and library restrictions. The developer's microservice is containerized, which would require App Engine Flexible, a VM-based environment that is not fully serverless and cannot scale to zero as cleanly. Thus, Cloud Run is the better choice.

Go deeper

Related to this question