A DevOps engineer must configure a production deployment pipeline. The AEM environment includes one author instance, two publish instances, and two dispatcher instances. A load balancer and CDN are also leveraged. In combination with load balancing, which step is required to make sure that the site is available during deployment?

A DevOps engineer must configure a production deployment pipeline. The AEM environment includes one author instance, two publish instances, and two dispatcher instances. A load balancer and CDN are also leveraged. In combination with load balancing, which step is required to make sure that the site is available during deployment?

• A. Point the CDN to the author instance during the publish instance deployment.
• B. Perform the deployment on one publish instance at a time.
• C. Disable the replication agents during the deployment.
• D. Clear the CDN cache after the author instance deployment.

Answer: (B)

The key idea is keeping the site available while updates are applied by using a rolling deployment strategy across publish instances behind a load balancer. With one author, two publish instances, and dispatchers/CDN in front, you don’t want to update both publish nodes at the same time. By deploying to one publish instance at a time, the load balancer keeps routing traffic to the healthy publish node that hasn’t been updated yet, so there’s always at least one instance serving requests. This minimizes downtime and avoids users hitting a partially updated or unavailable site. Pointing the CDN to the author would route read traffic to the author instance, which is not intended for production serving and could cause instability. Disabling replication during deployment can lead to content not being published to nodes correctly, and clearing the CDN cache after deploying the author doesn’t address availability during the deployment itself—it's about cache freshness, not ensuring continuous access.

The key idea is keeping the site available while updates are applied by using a rolling deployment strategy across publish instances behind a load balancer. With one author, two publish instances, and dispatchers/CDN in front, you don’t want to update both publish nodes at the same time. By deploying to one publish instance at a time, the load balancer keeps routing traffic to the healthy publish node that hasn’t been updated yet, so there’s always at least one instance serving requests. This minimizes downtime and avoids users hitting a partially updated or unavailable site.

Pointing the CDN to the author would route read traffic to the author instance, which is not intended for production serving and could cause instability. Disabling replication during deployment can lead to content not being published to nodes correctly, and clearing the CDN cache after deploying the author doesn’t address availability during the deployment itself—it's about cache freshness, not ensuring continuous access.