Cheat sheet & self-check

Pick an answer to see if you got it, and why.

1. Q1. What is the closest Google Cloud equivalent of an AWS account?

   1. AA folder
   2. BA project
   3. CA region
   4. DA VPC
   Show answer

   B. A project is the unit for resources, IAM, quotas, APIs and billing. Folders group projects, like organisational units group accounts.

   From lesson 01 · Google Cloud & GKE services and terms

2. Q2. In GKE, what plays the role of IRSA or EKS Pod Identity?

   1. ANode service account
   2. BWorkload Identity Federation for GKE
   3. CGoogle Groups for RBAC
   4. DBinary Authorization
   Show answer

   B. Workload Identity Federation lets a Kubernetes ServiceAccount act as an IAM principal, so pods get their own Google Cloud permissions instead of the node's.

   From lesson 01 · Google Cloud & GKE services and terms

3. Q3. Which statement about VPCs is true in Google Cloud but not in AWS?

   1. AA VPC belongs to one region
   2. BA VPC is global; its subnets are regional
   3. CSubnets are per zone
   4. DThere are no subnets
   Show answer

   B. A Google Cloud VPC spans all regions; each subnet lives in one region and covers all its zones. In AWS a VPC is regional and subnets are per AZ.

   From lesson 01 · Google Cloud & GKE services and terms

4. Q4. What is a node pool?

   1. AA group of clusters
   2. BA group of nodes in one cluster with the same configuration (machine type, image, labels)
   3. CA pool of IP addresses
   4. DA billing account
   Show answer

   B. Node pools are GKE's equivalent of managed node groups: same machine type, image, labels and taints, scaled and upgraded together.

   From lesson 01 · Google Cloud & GKE services and terms

5. Q5. In Autopilot, what do you pay for?

   1. AThe VMs you create
   2. BThe CPU, memory and storage your Pods request (plus a cluster fee)
   3. CNothing
   4. DOnly load balancers
   Show answer

   B. Autopilot bills per Pod resource requests; Google chooses and manages the nodes. Standard bills for the node VMs whether or not Pods use them.

   From lesson 02 · GKE architecture: Autopilot, Standard, regional and release channels

6. Q6. Why is a regional cluster the usual choice for production?

   1. AIt's free
   2. BIts control plane runs replicas in several zones, so a zone outage or a control-plane upgrade doesn't take the API away
   3. CIt has more features
   4. DIt's required for Autopilot only
   Show answer

   B. A zonal cluster has one control-plane replica: during its upgrade or a zone failure, kubectl and controllers can't reach the API (running Pods keep running).

   From lesson 02 · GKE architecture: Autopilot, Standard, regional and release channels

7. Q7. What does a release channel decide?

   1. AThe node machine type
   2. BWhich GKE versions the cluster gets and how soon after release, with automatic upgrades
   3. CThe VPC
   4. DThe billing account
   Show answer

   B. Rapid gets new versions first, Regular after more validation, Stable later still; Extended keeps a minor version for longer. GKE upgrades within the channel automatically.

   From lesson 02 · GKE architecture: Autopilot, Standard, regional and release channels