There is a property in the Tranquility Server configs called tasks.partitions. I’m wondering how exactly this is working. I have 3 tranquility servers consuming from a single kafka topic with 64 partition and sending these events to one datasource/task with two partitions. Is druid assuming that both partitions have different sets of data?
There is a partitionSpec for hadoop indexing tasks i think where you can specify partition strategy. Is there something like this for Tranquility Server?
All help is appreciated.