How do data prepper / dynamodb pipelines work?

Data Prepper
1

More of a few questions about how data prepper pipelines work (not sure if it matters that the source is dynamodb):

  1. Where is the state of each ingestion pipeline stored? is it in OpenSearch itself? is that something queriable?
  2. When updating a pipeline, by adding a new sink, is the process restarted so the new sink gets all the data? or does it just pick up from where it was and the other sinks don’t receive the old data?

thanks