This is a starting checklist you can customize as you write more detailed technical posts.
1. Understand the symptom
Is the job slow, expensive, failing, skewed, or producing too many small files?
2. Look at Spark UI
Start with stages, tasks, shuffle read/write, spill, executor time, and data skew.
3. Check Delta table layout
Review partitioning, file sizes, table statistics, clustering, and whether filters can skip data.
4. Fix one bottleneck at a time
Avoid random tuning. Change one thing, measure, and keep notes.