It seems that Prefect work pools are designed to coordinate deployments with workers through a single channel, similar to pub/sub systems. Each work pool can be optimized for specific infrastructure and workflow requirements, but the documentation does not explicitly mention targeting multiple node groups within a single work pool. You can control aspects of work pool behavior, such as concurrency, but it appears that each work pool is intended to manage a specific set of resources.
For more detailed information, you can refer to the
Prefect documentation on work pools.
If you are looking to manage multiple node groups, you might need to consider creating separate work pools for each node group and configure them individually to meet your scaling and granularity requirements.