Deployment Strategies: Canary vs. Blue-Green Rollouts

Answer key & explanations

Every question in this quiz, with the correct answer marked and an explanation of why it is right. Use it to revise before or after taking the quiz above.

1. 1.What is the foundational architectural difference between a Blue-Green deployment and a Canary deployment?

   • ABlue-Green runs on AWS; Canary runs on bare-metal infrastructure
   • BBlue-Green switches 100% of traffic from an existing idle environment to an updated one; Canary introduces changes to a small subset of traffic incrementally✓ Correct
   • CBlue-Green requires containerization; Canary cannot use containers
   • DCanary deployments do not allow for automated rollbacks

   Correct answer: Blue-Green switches 100% of traffic from an existing idle environment to an updated one; Canary introduces changes to a small subset of traffic incrementally

   Blue-Green provisions two full environments and cuts 100% of traffic over instantaneously; Canary exposes a new release to a small fraction of users first, gradually expanding traffic as metrics confirm stability.

2. 2.What is the primary resource drawback of maintaining a Blue-Green deployment architecture?

   • AIt requires double the infrastructure capacity during the deployment window to run both environments concurrently✓ Correct
   • BIt cannot support SSL/TLS certificate termination
   • CIt prevents databases from using relational schemas
   • DIt causes guaranteed downtime during the DNS propagation window

   Correct answer: It requires double the infrastructure capacity during the deployment window to run both environments concurrently

   Because Blue-Green maintains two identical environments simultaneously (Blue for live, Green for stage/next), it requires up to double the operational infrastructure during deploy cycles.

3. 3.In Canary deployments, what is an automated rollback trigger?

   • AA manual approval button clicked by a project manager in Slack
   • BA metric-driven policy (e.g., HTTP 5xx error rate spikes, latency breach) that halts traffic routing and reverts users to the stable baseline✓ Correct
   • CA scheduled timer that reverts every deployment after twenty-four hours
   • DAn operating system crash that uninstalls the Docker daemon

   Correct answer: A metric-driven policy (e.g., HTTP 5xx error rate spikes, latency breach) that halts traffic routing and reverts users to the stable baseline

   Automated canary systems (like Argo Rollouts or Flagger) monitor real-time SLIs; if error rates or latencies breach thresholds, they automatically pull traffic back to the stable version.

4. 4.What database schema migration pattern is essential to support zero-downtime Blue-Green deployments?

   • ADropping tables and recreating them with the new schema during traffic cutover
   • BThe Expand and Contract (Parallel Run) pattern, ensuring the database supports both the old and new application versions simultaneously✓ Correct
   • CLocking all database tables against writes until the deployment completes
   • DRunning two independent database instances that never synchronize

   Correct answer: The Expand and Contract (Parallel Run) pattern, ensuring the database supports both the old and new application versions simultaneously

   The Expand and Contract pattern introduces non-breaking additions (Expand) so both old and new code operate concurrently, only deprecating and cleaning up old columns (Contract) once rollouts succeed.

5. 5.How does a service mesh (like Istio or Linkerd) facilitate Canary deployments?

   • ABy recompiling container images dynamically in memory
   • BBy enabling fine-grained, weighted Layer 7 traffic routing rules (e.g., sending exactly 2% of HTTP requests to the canary deployment)✓ Correct
   • CBy automatically generating SQL migration scripts
   • DBy preventing Kubernetes nodes from using swap memory

   Correct answer: By enabling fine-grained, weighted Layer 7 traffic routing rules (e.g., sending exactly 2% of HTTP requests to the canary deployment)

   Service meshes use intelligent sidecar proxies to inspect and divide Layer 7 traffic by weight (e.g., 98% stable, 2% canary) or headers without changing physical Pod replica ratios.

6. 6.What is the role of Sticky Sessions during a Canary rollout?

   • APinning a specific user to either the baseline or canary version throughout their session to avoid inconsistent user experiences✓ Correct
   • BSaving all user database inputs to a persistent cache memory
   • CPreventing users from logging out of the application
   • DRestricting application traffic to users located in a single geographic city

   Correct answer: Pinning a specific user to either the baseline or canary version throughout their session to avoid inconsistent user experiences

   Sticky sessions (via cookies or session affinity) ensure that once a user lands on the canary version, they remain on it for the duration of their session, avoiding confusing UI flips.

7. 7.Why can relying solely on DNS record updates be problematic when switching traffic in a Blue-Green deployment?

   • ADNS servers do not support modern IPv6 addresses
   • BClient and ISP DNS caching ignores low TTL values, causing traffic to hit the old environment long after the switch✓ Correct
   • CDNS updates take down the physical network routers
   • DDNS records can only route traffic to web servers running Apache HTTPD

   Correct answer: Client and ISP DNS caching ignores low TTL values, causing traffic to hit the old environment long after the switch

   Client resolvers and ISP caching frequently ignore low Time-To-Live (TTL) values, meaning traffic will trickle to the deprecated environment until upstream caches expire.

8. 8.What is Dark Launching in canary release terminology?

   • ADeploying software changes late at night during non-business hours
   • BReleasing features to production that execute in the background or mirror real traffic without exposing user-facing UI changes✓ Correct
   • CShutting down all monitoring dashboards during a deployment
   • DDeploying software on servers located in underground data centers

   Correct answer: Releasing features to production that execute in the background or mirror real traffic without exposing user-facing UI changes

   Dark launching deploys features to production quietly—often by shadowing live production traffic or hiding UI triggers behind feature flags—to test backend scalability without user impact.

9. 9.What is Shadow Traffic (or Traffic Mirroring) during canary testing?

   • AStealing user credentials over unencrypted Wi-Fi networks
   • BDuplicating incoming live production requests and sending a copy asynchronously to a test service, ignoring its response✓ Correct
   • CRouting internal developer traffic through public VPNs
   • DHiding the destination IP address of production servers

   Correct answer: Duplicating incoming live production requests and sending a copy asynchronously to a test service, ignoring its response

   Traffic shadowing duplicates real production requests and routes them to the new deployment out-of-band; responses are discarded, allowing realistic load testing without user risk.

10. 10.What is the difference between a Rolling Update in standard Kubernetes and a true Canary rollout?

    • ARolling updates replace Pods incrementally without intelligent health metric validation; Canary evaluates live traffic metrics before widening distribution✓ Correct
    • BRolling updates only work on single-node clusters; Canary requires multiple cloud regions
    • CRolling updates cannot be stopped once initiated
    • DCanary rollouts can only be configured using raw Bash scripts

    Correct answer: Rolling updates replace Pods incrementally without intelligent health metric validation; Canary evaluates live traffic metrics before widening distribution

    Standard rolling updates replace old Pods with new ones based purely on container readiness probes, lacking the canary ability to analyze real user traffic metrics (error rates, latencies) before proceeding.

11. 11.What is a Feature Flag (Feature Toggle), and how does it complement Canary deployments?

    • AA tool that compiles code into binary flags
    • BA software mechanism that decouples code deployment from feature release, enabling dynamic activation or deactivation of features at runtime✓ Correct
    • CA compiler directive that deletes untested functions during build time
    • DA hardware switch located on physical rack servers

    Correct answer: A software mechanism that decouples code deployment from feature release, enabling dynamic activation or deactivation of features at runtime

    Feature flags decouple code deployment from feature exposure; code can be safely released to production dark and progressively enabled for target segments without redeployment.

12. 12.What constitutes an A/B test in contrast to a standard technical Canary rollout?

    • AA/B tests evaluate business metrics and user behavior between two variants; Canary rollouts validate system health, stability, and error rates✓ Correct
    • BA/B tests are used exclusively by database administrators; Canary rollouts are used by frontend developers
    • CA/B tests require running two distinct operating systems simultaneously
    • DThere is no difference; they are identical terms for the exact same process

    Correct answer: A/B tests evaluate business metrics and user behavior between two variants; Canary rollouts validate system health, stability, and error rates

    While both route subsets of users to different versions, a canary deployment tests technical stability and performance, whereas an A/B test measures business metrics and user conversion.

13. 13.What happens in a Blue-Green deployment if an application requires persistent, non-distributed local file storage?

    • AThe application will corrupt data unless files are offloaded to centralized shared storage (like an S3 bucket or NFS mount)✓ Correct
    • BThe file system automatically replicates across local hard drives in under one millisecond
    • CThe operating system prevents the Blue environment from cutting over
    • DThe deployment process converts local text files into relational database tables

    Correct answer: The application will corrupt data unless files are offloaded to centralized shared storage (like an S3 bucket or NFS mount)

    If an app writes state to local disk, the two independent environments will drift; zero-downtime cutovers require state and file storage to be externalized to shared storage.

14. 14.What is Progressive Delivery?

    • AShipping software exclusively on physical flash drives via postal mail
    • BAn umbrella term combining continuous delivery, canary deployments, feature flags, and observability-driven automated rollouts✓ Correct
    • CDeploying software updates line by line to live servers
    • DA methodology where only senior staff are allowed to push code

    Correct answer: An umbrella term combining continuous delivery, canary deployments, feature flags, and observability-driven automated rollouts

    Progressive Delivery builds on Continuous Delivery by using feature flags, traffic shifting, canary analyses, and automated rollbacks to roll out changes safely with minimal blast radius.

15. 15.How does Header-based Routing assist internal canary testing?

    • AIt encrypts the HTTP request headers using RSA keys
    • BIt routes requests to the canary deployment only if a specific HTTP header (e.g., 'X-Canary: true' or a staff session cookie) is present✓ Correct
    • CIt strips out all user cookies before hitting the backend microservices
    • DIt limits HTTP request sizes to 512 bytes

    Correct answer: It routes requests to the canary deployment only if a specific HTTP header (e.g., 'X-Canary: true' or a staff session cookie) is present

    Header routing allows operators or internal employees to test the canary build directly in production by matching on specific headers or cookies while keeping public traffic on the stable version.

16. 16.In a canary analysis, what is the 'Baseline' deployment?

    • AA clean server running an empty operating system
    • BAn instance running the currently stable production version, deployed under identical conditions to provide a fair statistical comparison against the Canary✓ Correct
    • CA backup database containing synthetic mock data
    • DThe initial git commit made when the project repository was created

    Correct answer: An instance running the currently stable production version, deployed under identical conditions to provide a fair statistical comparison against the Canary

    To eliminate noise from long-running process artifacts, automated canary analysis often deploys a fresh 'Baseline' alongside the 'Canary' with identical resources to make an accurate comparison.

More free quizzes

• ITDistributed Systems: Consensus & Fault ToleranceDeep dive into distributed systems engineering, covering consensus protocols (Raft/Paxos), the CAP theorem, and vector clocks.16 questions
• ITLLM Engineering: Fine-Tuning, Alignment & ServingTest your knowledge of transformer architectures, PEFT techniques (LoRA), RLHF, vLLM, and production deployment strategies.16 questions
• ITApplied AI: Systems, Embeddings & RAGAssess your understanding of practical machine learning architectures, embedding spaces, and Retrieval-Augmented Generation systems.16 questions