01. An operations lead notes that one release strategy briefly requires roughly double the normal compute footprint because a complete second environment runs in parallel before traffic is switched. Which strategy is being described, with its main trade-off?
a) Blue-green deployment, giving fast rollback but temporarily doubling resource use
b) Rolling deployment, replacing nodes gradually with little spare capacity
c) Canary deployment, exposing a small user subset to limit risk
d) Recreate deployment, stopping the old version before starting the new
02. When an order is placed, an order service must notify several independent downstream services (inventory, billing, and shipping) at the same time. The order service should not need to know which services consume the notification. Which characteristics describe the publish/subscribe pattern that fits this need?
(Choose two.)
a) Each message is consumed by exactly one worker and then removed
b) The publisher is decoupled from consumers and does not address them directly
c) The publisher must maintain a direct connection to each subscriber
d) Delivery is strictly synchronous request-response between two endpoints
e) A single published message is delivered to every subscribed consumer
03. Before touching production, an engineer wants to preview exactly which resources an IaC change will add, modify, or destroy, without enacting anything yet. Which step provides this preview?
a) Running drift detection after deployment
b) Running apply to enact the changes immediately
c) Running plan to generate an execution preview
d) Committing the code to version control
04. Which statements correctly describe a Type 1 (bare-metal) hypervisor?
(Choose two.)
a) It is the typical choice for production data-center and cloud virtualization
b) It runs as an application installed on a desktop host operating system
c) It shares the host OS kernel with its workloads instead of virtualizing hardware
d) It can host only a single virtual machine per physical server
e) It runs directly on the physical hardware without a host operating system
05. During a migration wave, one legacy component has an unresolved data-residency constraint and a specialized hardware dependency that cannot be satisfied in the cloud yet. The team decides to leave it running on-premises for now and revisit it in a later wave while migrating everything else.
Which migration decision is being applied to that component?
a) Rehost the component onto cloud virtual machines immediately
b) Retire the component because it is no longer needed
c) Refactor the component before the rest of the workload moves
d) Retain the component in place and revisit it in a later wave
06. A large team keeps each feature on its own long-lived branch that lives for weeks. Merges back to main have become large and painful, and integration bugs are discovered late. The team wants to integrate small changes frequently and shrink merge conflicts.
Which branching approach BEST addresses these problems?
a) Keep each feature on a dedicated long-lived branch and merge only at release
b) Adopt trunk-based development, committing small changes frequently to a shared main behind feature flags
c) Freeze the main branch and forbid commits until the release date
d) Give each developer a private fork that is never merged back
07. A workload still protects its data with a single encryption key that was created years ago and has never changed. A security review flags this as a risk. Which two practices most directly reduce the risk tied to this long-lived key?
(Choose two.)
a) Configure the key never to expire to avoid operational disruption
b) Store the key in application source code so services can read it easily
c) Enable scheduled automatic rotation of the key
d) Adopt envelope encryption so the master key can rotate without re-encrypting all data
e) Reuse the single key across every workload and environment
08. Users report slowness. Dashboards show request rate and error rate near normal, but a queue-depth and memory-utilization signal is climbing steadily toward its limit. Which of the four golden signals is indicating the problem?
a) Saturation
b) Traffic
c) Errors
d) Latency
09. A tier-1 transactional database must minimize both the amount of data lost and the time to recover after a zone failure. Which two configurations BEST meet these goals together?
(Choose two.)
a) A manually activated cold standby site
b) Automated failover to a running hot standby
c) Nightly full backups exported to object storage
d) Synchronous replication to a standby in another zone
e) Extending backup retention to seven years
10. Two engineers run apply against the same shared IaC configuration at the same moment. Which mechanism prevents the concurrent operations from corrupting the recorded infrastructure state?
a) Drift detection
b) Branch protection
c) State locking
d) Idempotency