Control SRA batching
The Google Publisher Tag (GPT) library's Single Request Architecture (SRA) mode batches multiple ad slots into a single ad request. This provides performance benefits, but is also necessary to guarantee that competitive exclusions and roadblocks are honored.
By default, all ad slots defined prior to calling display() or refresh()
will be batched and requested together when SRA mode is active. There are some
situations where you may want more control over this batching behavior, however.
For example, to prioritize the loading of specific slots or
to avoid exceeding the
maximum number of slots in a single SRA request.
This example demonstrates how to apply techniques for controlling ad loading and refresh to SRA requests.