I read from the following article: https://databricks.com/blog/2015/07/30/diving-into-spark-streamings-execution-model.html
which says that Spark Streaming has a future direction for "Event time and out-of-order". I am wondering if it is supported now. I my scenario, I have three streams, stream1, stream2, stream3. Each stream has its own timestamp. I want to join stream1, stream2 and stream3 every 5 minutes. Sometimes stream2 will be delayed(hours or even days)...I am wondering if Spark Streaming is a good fit for my scenario? Thanks, Yifei