Associate Cloud Engineer Practice Test

What this test covers

One concept per question, across the topics Associate Cloud Engineer draws on.

• Minimum storage duration makes cold classes expensive for short-lived data
• Match storage class to actual access frequency, not to a general sense of importance
• Stopping an instance halts compute charges but not all associated resource charges
• Spot capacity suits fault-tolerant, interruptible work
• IAM is additive and inherits downward; lower levels cannot subtract an inherited grant
• Organisation policy is a preventative control that binds even project administrators
• Billing accounts control invoicing; the resource hierarchy and labels provide cost attribution
• Global HTTP(S) traffic with one anycast address means the global Application Load Balancer
• High availability provides automatic failover; read replicas do not
• Cloud NAT provides outbound-only internet access for instances without public addresses
• A locked retention policy is immutable and binds even administrators
• Signed URLs grant time-limited access to a specific object without an identity
• Commit to the baseline, absorb variability with flexible capacity
• IAM conditions can bind access to time, resource names or other attributes
• Organisation policy provides preventative guardrails that inherit down the hierarchy
• Sole tenancy provides dedicated physical hardware for licensing and compliance
• Baking a custom image minimises startup time and makes instances reproducible
• Zones are independent failure domains within a region
• Bucket names share a single global namespace
• Log-based metrics turn log content into numeric time series you can alert on
• Autohealing recreates unhealthy instances; load balancer health checks only redirect traffic
• Shared responsibility: the provider secures the infrastructure, the customer secures their use of it
• Firewall rules are evaluated by priority, lowest number first, and the first match wins
• Edge caching reduces both origin load and egress cost simultaneously
• Pod scaling and node scaling are separate mechanisms
• Pending pods indicate a scheduling problem, almost always insufficient capacity
• Local SSD trades durability for maximum performance
• Grant to groups so access follows role changes automatically
• Scale to zero suits spiky traffic with a small operations team
• Shared read-write access with file semantics requires network file storage
• Implied firewall rules deny all ingress and allow all egress
• Least privilege is both the narrowest role and the narrowest scope