Expand my Community achievements bar.

Guidelines for the Responsible Use of Generative AI in the Experience Cloud Community.

Replication agent batch mode and trigger size

Avatar

Level 1

What would you suggest as "Trigger Size" value for a replication agent in batch mode?

I understand that it should be a balance between not being too small (so package get created and replicated continuously, but not too big (because create and replicate big packages can also create problems).

 

What would you suggest as a balanced value? 50 mb, 100 mb, more?

Topics

Topics help categorize Community content and increase your ability to discover relevant content.

2 Replies

Avatar

Community Advisor

Hi, 

 

It's difficult to determine an exact value because it depends on the type of replication activities occurring in the AEM instance, which can vary between environments. Additionally, it depends on the "max-time" value you set, as replication will trigger based on either the "trigger size" or the "max-time." If you set a low "max-time" (the default is 2 seconds), then the "trigger size" may not significantly impact the process. However, if you set the "max-time" to 1 hour, the "trigger size" becomes critical.

That being said, my suggestion would be to quantify the size of your pages or assets rather than focusing on a specific value in MB. Estimate how many pages or assets you want to wait for before replication is complete. This will allow you to set a more practical figure based on the amount of content, rather than just focusing on the size in MB. For example, if you want to wait until 8 pages are ready and each page is approximately 250 KB, you can calculate the "trigger size" accordingly. This approach can help you determine a more accurate and applicable value based on your needs.


Hope this helps.



Esteban Bustamante

Avatar

Level 1

Is the trigger size defined in KB or MB ?so whenever the size reaches, it batches and replicates immediately or it wait for all the pages to be batched independently  and replicates all batches together ? How efficient is this batch replication in comparison with normal bulk replication ?