On which instance should a flush agent be configured to prevent invalidation timing issues after invalidation?

On which instance should a flush agent be configured to prevent invalidation timing issues after invalidation?

• A. Dispatcher
• B. Publish
• C. Author
• D. Load-balancer

Answer: (B)

When content is changed, the cache must be refreshed in the same place that serves the live content to users. In an AEM setup, end users request come from the Publish environment through the Dispatcher cache. Configuring a flush (invalidate) mechanism on the Publish instance ensures that once content is published, the Dispatcher cache is invalidated at the exact point the new content becomes available, so visitors consistently see the latest version. If you place the flush tool on the Author, the invalidation can occur before the content actually exists on Publish, leading to timing gaps where stale content might be served. Placing it on the Publish side aligns the invalidation with the actual delivery path, and if there are multiple Publish nodes, you can ensure each one handles its own cache properly. Dispatcher or load-balancer-level flush configurations don’t synchronize with the publication flow as reliably, which is why configuring the flush agent on the Publish instance is the best choice.

When content is changed, the cache must be refreshed in the same place that serves the live content to users. In an AEM setup, end users request come from the Publish environment through the Dispatcher cache. Configuring a flush (invalidate) mechanism on the Publish instance ensures that once content is published, the Dispatcher cache is invalidated at the exact point the new content becomes available, so visitors consistently see the latest version. If you place the flush tool on the Author, the invalidation can occur before the content actually exists on Publish, leading to timing gaps where stale content might be served. Placing it on the Publish side aligns the invalidation with the actual delivery path, and if there are multiple Publish nodes, you can ensure each one handles its own cache properly. Dispatcher or load-balancer-level flush configurations don’t synchronize with the publication flow as reliably, which is why configuring the flush agent on the Publish instance is the best choice.