it is the default and only driver. But you should also take a look at this part:
This allows Vapor to ensure the correct number of failed attempts is accessible by your application, even if a queued job does not complete due to exceeding the maximum execution time enforced by Lambda (15 minutes)
so I am not sure how good of an approach is to have a serverless application that should handle long running jobs.
Actually this "red" part (right in the introduction) is the proper answer for your question:
Queued Job Time Limits
Currently, serverless applications on AWS may only process a single request (web or queue) for a maximum of 15 minutes. If your queued jobs take longer than 15 minutes, you will need to either chunk your job's work into smaller pieces or consider another deployment solution for your application. In addition, a queued job may not have a "delay" greater than 15 minutes.