DOP-C02 Practice Question 52: Beanstalk Immutable Deployment
AWS Certified DevOps Engineer – Professional (DOP-C02) · Test 6 · question 2 of 10
A team deploys a new application version to a production Elastic Beanstalk environment several times a week. They want any deployment that fails its health checks to roll back automatically without any downtime, and they want the full capacity of the original, known-good instances to keep serving traffic until the new instances have proven healthy. A short period of double capacity during the deployment is acceptable.
Which Elastic Beanstalk deployment policy meets these requirements?
Show answer and explanation
Answer:
Immutable deployments launch a second, temporary Auto Scaling group behind the environment's load balancer with the new application version. Elastic Beanstalk starts one instance there, waits for it to pass health checks, then launches the rest to match the original group's capacity. Only after all the new instances are healthy does it move them into the original Auto Scaling group and terminate the old instances. If the new instances never pass health checks, Elastic Beanstalk simply terminates the temporary Auto Scaling group, which requires no traffic shift back and leaves the original instances untouched the whole time, giving a zero-downtime rollback at the cost of briefly doubling capacity.
Option A, plain Rolling, updates instances in the existing Auto Scaling group in batches, so capacity briefly drops during the update and a bad batch can already be serving live traffic before it fails health checks.
Option B, Rolling with additional batch, avoids the capacity dip but still updates the existing Auto Scaling group's instances directly, so a failed batch still needs to be rolled back in place rather than simply discarded.
Option C, All at once, updates every instance simultaneously with no new instances held in reserve, causing downtime if the new version fails.
Reference: Immutable environment updates