What are the four main compute layers?
IaaS (Compute Engine), CaaS (GKE), PaaS (App Engine), FaaS (Cloud Run/Functions).
When use Compute Engine?
Lift-and-shift legacy apps, full OS control, custom VM tuning.
What are Managed Instance Groups (MIGs)?
Groups of identical VMs with autoscaling and auto-healing.
What’s the difference between GKE Standard and Autopilot?
Standard = user manages nodes; Autopilot = Google manages infrastructure.
When to use App Engine Standard vs Flexible?
Standard = fully managed, scales to zero; Flexible = custom runtimes, more control.
Cloud Run vs Cloud Functions?
Cloud Run = containerized services, HTTP-based; Functions = single event triggers.
What is Anthos used for?
Hybrid/multi-cloud management and policy enforcement for GKE clusters.
What are preemptible VMs?
Short-lived (24h max), cost-efficient VMs for batch/stateless jobs.
How to ensure VM fleet auto-heals and scales?
Managed Instance Groups + health checks.
Exam clue: “Run container without managing servers.”
Cloud Run.
Exam clue: “Event-triggered automation.”
Cloud Functions.
Exam clue: “Hybrid K8s across on-prem and cloud.”
Anthos.
How do you lower compute costs for long-term workloads?
Use committed use discounts or sustained usage discounts.
How do you lower costs for transient jobs?
Preemptible/spot instances.
How do instance templates assist scaling?
Define uniform VM configuration for MIGs and auto-scaling policies.
What service builds and stores container images?
Cloud Build + Artifact Registry.
What service automates deployment to GKE/Run?
Cloud Deploy.
Exam cue: “Autoscale containers, limit ops overhead.”
GKE Autopilot or Cloud Run (depending on context).
Exam cue: “Web app with version rollouts, zero ops.”
App Engine Standard.
Exam cue: “Legacy DB with full OS control.”
Compute Engine VM.