Why this is in my collection

From the publisher:

Quickly detect, troubleshoot, and prevent a wide range of data issues through data observability, a set of best practices that enable data teams to gain greater visibility of data and its usage. If you're a data engineer, data architect, or machine learning engineer, or if the quality of your work depends on the quality of your data, this book shows you how to focus on the practical aspects of introducing data observability in your everyday work.

Highlights

  • Observability is built in, monitoring is bolted on — Petrella's core distinction: visibility into data and its usage must be produced by the systems themselves, not reconstructed afterwards by an external watcher; instrumentation as a structural property, not an add-on product.
  • Expectations as code — declaring what data should look like (shape, freshness, distributions) and executing those declarations continuously turns quality from opinion into rules; the rules-as-third-pillar argument applied at runtime.
  • Shift observability to the producer — the engineers who generate data should instrument it at the source, rather than downstream teams inferring problems from symptoms; the same shift-left economics as fixing ambiguity at requirements time.
  • Lineage is the runtime shadow of the model — knowing where data came from and what consumed it is model information observed in motion; a warehouse generated from a model gets this lineage by construction instead of by archaeology.
  • Detection complements prevention — and checks can be generated — observability finds the issues that slip through, while a sound model prevents whole classes of them; through the MDDE lens the natural synthesis is deriving the observability checks from the same model that generates the pipeline.
  • Silent data issues compound downstream — the longer a broken assumption travels, the more expensive it becomes; the cost curve that justifies structural guarantees early over heroic debugging late.

Highlights on this page are generated with the help of AI.