A Horror Story told through the view of building an ETL pipeline, and how i got it super-wrong - before then getting it right. Showing code that replicates the issues i had at the time, and how to fix it - looking at concurrency, asynchronicity, optimization, resilience and persistence