Roadmap
Features
Below is a non-exhaustive list of potential features for future releases (presented in no particular order):
- Provide state-of-the-art stream processing autoscaling algorithms out of the box. A good first candidate is DS2.
- Provide simple autoscaling algorithms (e.g., HPA) out of the box.
- Support additional native-Kubernetes Resources that implement the Scale Sub-resource. This would allow users to scale any native Kubernets framework, e.g., Numaflow.
- Integrate scale strategies into the KTA Kubernetes Operator (cf. HPA workload stabilization).
- Monitor changes in stream processing application and reconcile state back to the desired state of KTA.
Tasks
Below is a non-exhaustive list of potential tasks (work items that do not introduce a new feature but, e.g., improve user experience) for future releases (presented in no particular order):
- Publish the KTA Python SDK to PyPI.
- Provide helm charts.
- Improve error handling (robustness).
- Simplify Flink deployment (e.g., automated job graph retrieval from task manager).
- Refine SDK and API design.
- Introduce website versioning.