Hello, We are looking into running batch jobs on Flink clusters. Intuitively, Session Clusters seem like an excellent deployment mode.
However, the challenge is that batch jobs may have different image requirements, especially for ML workloads. Currently, task managers must be homogeneous, meaning you would need a massive Docker image that needs to be updated whenever you have different image requirements. What are your thoughts on offering a Session cluster that supports heterogeneous task managers? It would mean that the scheduler would need some level of job isolation and task manager image discovery. It would also mean that the operator/api would need the ability to pass different images with the job. Thanks, Thanks, Ryan van Huuksloot Sr. Production Engineer | Streaming Platform [image: Shopify] <https://www.shopify.com/?utm_medium=salessignatures&utm_source=hs_email>
