Hello @jkff and @Graham: I am having dataflow streaming pipeline, which uses WriteToBigQuery after reading from original source. I have similar requirement where I want to achieve two things: 1. handling WriteToBigQuery error records and 2. execute big query SQL which has to be executed only after WriteToBigQuery. For 2nd requirement: Approach mentioned by @jkff wont work for me as it will submit separate dataflow steaming job which will be async with original job. I have question about Approach mentioned by @Graham, If I do that way, will that get executed everytime there is streaming data coming from original source?