The Databricks pipeline that ran green for a year and was wrong the whole time
The Databricks Data EngineerShow Notes
A revenue-attribution pipeline runs overnight, bronze to silver to gold, and lands a clean set of numbers the whole company uses to decide where the money goes. Every morning: green. Every morning, everyone moves on with their day. For a year, nobody looked any closer.
The job succeeded every single night, so everyone trusted the data. But "finished" and "correct" are two different words, and the space between them had been quietly swallowing a slice of revenue for eleven months, with no error, no warning, and not one alert.
In this episode: - Why a Databricks job can run green every day and still feed wrong numbers into decisions that matter - The one question that separates "did it run" from "is it right" and cracks silent failures wide open - How an ordinary upstream schema change and an inner join can drop rows for months without tripping anything - The single correctness check that would have caught this on night one instead of month eleven - What to say in the room after an incident like this if you want leadership to trust you with the next hard problem
This episode is for Databricks data engineers who own pipelines whose numbers someone actually makes decisions on. Whether you've ever said "the pipeline's fine, it's probably the dashboard" or you're staring at an all-green run history right now, you'll walk away knowing exactly what to check before you trust it again.
--- Helping 18,000+ Databricks data engineers become seniors: interview like seniors, execute like seniors, think like seniors.
Follow The Databricks Data Engineer for new episodes every Monday, Wednesday, and Friday.
LinkedIn: linkedin.com/in/jrlasak Newsletter: dataengineer.wiki
#DataEngineering #Databricks #DataEngineer #CareerGrowth #ApacheSpark #DeltaLake
More from The Databricks Data Engineer
The rollout is done. Every production table migrated to Unity Catalog, every one owned by a team group, and the revenue table carries a certified check mark. The platform lead posts in Slack: fully governed. Monday morning, finance replies in the thread. The revenue on the finance dashboard and the revenue in the sales deck are about three percent apart. Which one's real?
Two proposals sit on a head of data's desk, both about the nightly pipeline that feeds the finance dashboard. One swaps the full reload for incremental loads with Auto Loader: it finishes in a fraction of the time and it's cheaper every month after the first. The other keeps the full reload, puts it on a bigger job cluster and starts it an hour earlier. It costs more. Forever. She asks one question, "What happens if this breaks on quarter close?", and approves the second one.
Your nightly Databricks job ran long, so you open the Spark UI. Two hundred tasks. One hundred ninety-nine finished in about a minute. One is still running, and it's been running for forty. You double the cluster and rerun it. Same forty minutes. Nearly twice the bill.
Sam is four years into a Databricks role, up for senior this cycle. Merged pull requests on the ingestion layer. The nightly job he rescued at three in the morning when the orders table stopped landing. Then his manager asks for one call where the answer wasn't obvious, he framed it, and he lived with the result. Sam starts three different sentences. None of them finish.
Open the workspace. Start a cluster. Get coffee. Come back, and sometimes it's ready. If you've been doing this a few years you can feel that wait in your shoulders, and it's the least interesting part of the story.