Answer: Rolling update

## Explanation **Correct Answer: A) Rolling update** **Why Rolling Update is the correct choice:** 1. **Maintains capacity**: Rolling updates deploy new versions to a Managed Instance Group (MIG) by gradually replacing instances while maintaining the desired capacity throughout the process. 2. **Zero-downtime deployment**: Rolling updates work by: - Creating new instances with the updated template - Adding them to the MIG - Removing old instances - This happens in batches to ensure the MIG always has the required number of healthy instances 3. **Controlled deployment**: You can configure parameters like: - Maximum number of instances being updated at once - Minimum number of healthy instances during the update - Instance replacement method (proactive or opportunistic) **Why other options are incorrect:** - **B) Recreate instances**: This would terminate all instances at once and create new ones, causing downtime and reducing capacity during the transition. - **C) Manual replacement**: This is inefficient and error-prone, requiring manual intervention for each instance replacement. - **D) Stop autoscaling**: This doesn't help with version deployment; it only stops the autoscaling feature but doesn't provide a mechanism for updating instances to a new version. **Key takeaway**: When deploying new versions to a Managed Instance Group while maintaining capacity and minimizing downtime, rolling updates are the recommended Google Cloud approach.

Author: Rodrigo Sales