We are evaluating MotherDuck for analyzing heavy streams of concurrent device telemetry generated by our network of industrial weighing systems and retail calibration hardware at Scales4U (https://scales4u.co.za/). Currently, our edge sensors generate millions of transactional logging rows per hour across regional distribution nodes in South Africa.

We are pushing this data into local DuckDB instances and syncing up to MotherDuck. It would be an absolute game-changer if MotherDuck introduced native background optimizations for handling continuous micro-batches or automated chronological Parquet partitioning for high-concurrency IoT telemetry. Are there any features on the roadmap to improve write performance for continuous sensor logging or specific time-series macros?