Objective: name the control you need
Use the responsibility boundary from earlier lessons to compare compute options against a concrete workload. The relevant question is which operating control is necessary and which operating work the team can accept. A requirement for a specific guest operating system differs from a requirement for Kubernetes orchestration, which differs from simply needing an HTTPS application endpoint. All three can involve application code, but they do not justify the same platform. Keep compatibility, staffing, and delivery constraints visible while comparing alternatives.
Three starting points
Compute Engine provides self-managed virtual machines and bare metal instances. A VM can support an existing workload that needs guest-OS control, with corresponding customer software responsibilities. GKE is managed Kubernetes for deploying and operating containerized applications; it is relevant when Kubernetes capabilities or configuration are required. Cloud Run is a managed application platform for code, functions, and containers that avoids creating a cluster merely to run an application. These distinctions establish candidates, not a universal cost or performance ranking. Verify the selected service’s actual workload requirements.
Execution shape matters
Cloud Run documentation distinguishes services, jobs, worker pools, and instances. A service handles requests, while a job runs tasks to completion; other modes address different background or long-lived execution patterns. Do not apply a statement about request-based service scaling to every Cloud Run mode. Persistent business data also requires deliberate storage: a disposable container instance’s local writable file system is not a durable records system. Identify what must survive a restart before calling a workload ready to migrate.
Worked example
Fictional case: a surveying firm has a stable Windows application with a vendor-required guest-OS component. It needs to leave an aging server room before its lease ends. First verify the compatibility requirement and the deadline. A Compute Engine VM is a reasonable candidate to investigate because guest-OS control is explicit. Second, inventory dependencies, data, credentials, and the operating tasks the company retains. Third, test the workflow in the candidate environment and plan recovery if cutover fails. Calling this a lift-and-shift does not show that the application has been redesigned or that its operating burden has disappeared.
Changed condition: a new bounded task
The same firm now builds a separate containerized report generator that starts from a supplied input, writes the finished output to durable storage, and exits. It has no Kubernetes-specific requirement. Investigate a Cloud Run job before assuming that a Kubernetes cluster is necessary. If a later requirement genuinely needs Kubernetes orchestration or infrastructure configuration, evaluate GKE on that basis. A migration can therefore contain different decisions for different components. Rewriting the original Windows application would be a separate change with its own testing and schedule risk.
Transfer check
A developer proposes a Cloud Run service because a nightly script should run to completion and stop. What should be reconsidered? Check: examine the job execution mode against that bounded task rather than treating every container as an HTTP service. Verify input, output persistence, failure handling, and the actual schedule separately. Retrieve the VM, Kubernetes, and run-to-completion distinctions using the cards. Next, study API boundaries and portability: changing a runtime does not automatically make dependencies interchangeable.
Try it, then explain it.
Complete a short set without looking at the example. After each answer, explain which evidence or rule supports your decision and why the alternatives do not fit.
Sources & authorship
Original OpenExam explanations and examples. References support the concepts; they are not the source of copied questions. Independent educator review is pending.
Shared responsibilities and shared fate on Google Cloud | Cloud Architecture Center | Google Cloud DocumentationCompute Engine overview | Google Cloud DocumentationGKE overview | Google Kubernetes Engine (GKE) | Google Cloud DocumentationWhat is Cloud Run | Google Cloud Documentation