DOP-C02 Configuration Management and IaC Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Blue/green deployment with a fixed number of instances.

  Why it's wrong here

  Blue/green deployment with a fixed number of instances replaces the entire environment at once; traffic is cut over from the old environment to the new environment, so the blast radius is 100% of the fleet, not a small percentage. Even if the new environment has the same fixed instance count, you cannot gradually shift a small percentage of traffic instance-by-instance without implementing additional weighted routing, which is not described here. It also doubles infrastructure cost during the deployment and is overkill for a change that needs a controlled, small rollout.

• ✗

  In-place deployment with 'CodeDeployDefault.AllAtOnce' configuration.

  Why it's wrong here

  In-place deployment with 'CodeDeployDefault.AllAtOnce' configuration instructs CodeDeploy to deploy to all instances in the deployment group simultaneously, meaning every instance is updated in the same deployment event. This maximizes the blast radius and provides no opportunity to validate the new revision on a small subset before the whole fleet is affected. It directly violates the requirement to rollout to a small percentage of instances at a time, as there is no staged progression or health-check pause between instances.

• ✗

  In-place deployment with 'CodeDeployDefault.HalfAtATime' configuration.

  Why it's wrong here

  In-place deployment with 'CodeDeployDefault.HalfAtATime' configuration divides the fleet into two halves and deploys to the first half, waits for health checks, then deploys to the second half. The step size is fixed at 50% of the instances in the deployment group, which is far too coarse for a requirement to update a small percentage at a time. If the target small percentage is, say, 5-10%, a 50% stage gives no protection against a bad revision affecting half of the production traffic before rollback can trigger.

• ✓

  In-place deployment with 'CodeDeployDefault.OneAtATime' configuration and automatic rollback enabled.

  Why this is correct

  In-place deployment with 'CodeDeployDefault.OneAtATime' configuration is the built-in CodeDeploy strategy that stages the deployment to a single instance at a time, waiting for the instance to pass health checks before proceeding to the next. When the fleet size is reasonably large, one instance represents a small percentage of total traffic, satisfying the gradual rollout requirement. Enabling automatic rollback ensures that if any instance fails its health check or a deployment hook returns a non-zero exit code, CodeDeploy immediately redeploys the previous revision and stops the deployment, minimizing impact.

Go deeper

Related to this question