As the title suggests, the spark library is too slow for our use case. I understand a physical partition maxes out at 10k RU/s, but even when I over provision my RU/s, it still takes a long time to load data into cosmosDB. Is there a way to force cosmosDB to be more aggressive when splitting data across physical partitions? This is ahuge bottleneck for my team
As the title suggests, the spark library is too slow for our use case. I understand a physical partition maxes out at 10k RU/s, but even when I over provision my RU/s, it still takes a long time to load data into cosmosDB. Is there a way to force cosmosDB to be more aggressive when splitting data across physical partitions? This is ahuge bottleneck for my team