SAP performance tuning · LessonBy Anitha M, SAP Trainer, 13 yrs · Published · SAP S/4HANA 2023 · all levels
Performance tuning
Performance tuning keeps SAP fast and scalable by finding and fixing bottlenecks across the database, application server and custom code. It is a measurement-driven discipline, never guesswork.
Quick answer
Always measure first: baseline with ST03, drill to the worst offenders, trace the cause, fix one thing (usually bad SQL/code or an index), and re-measure.
Key takeaways
- Database: expensive SQL, missing indexes, outdated statistics (usually the first bottleneck).
- Custom ABAP: SELECT-in-loop, SELECT *, nested loops on big tables.
- Buffers/memory: under-sized program/table buffers.
- Watch out: Tuning without measuring.
Where performance is won or lost
- Database: expensive SQL, missing indexes, outdated statistics (usually the first bottleneck).
- Custom ABAP: SELECT-in-loop, SELECT *, nested loops on big tables.
- Buffers/memory: under-sized program/table buffers.
- Work processes: wrong mix/number for the workload.
The tuning toolkit
- ST03 workload analysis.
- ST05 SQL trace; SAT/ST12 runtime analysis.
- ST04 / HANA plan analysis.
- SM50/SM66 work-process monitoring.
The method
Always measure first: baseline with ST03, drill to the worst offenders, trace the cause, fix one thing (usually bad SQL/code or an index), and re-measure. On HANA, push logic to the database and avoid row-by-row processing.
Common pitfalls
- Tuning without measuring.
- Ignoring custom-code SQL.
- Under-sized DB/buffers.
Practice challenge
+0 XPStreak ×0
Question 1 of 3
Which statement is true of Performance tuning?
Frequently asked questions
What does the term Performance tuning refer to in SAP?
Performance tuning keeps SAP fast and scalable by finding and fixing bottlenecks across the database, application server and custom code. It is a measurement-driven discipline, never guesswork.
What else is worth knowing about Performance tuning?
Always measure first: baseline with ST03, drill to the worst offenders, trace the cause, fix one thing (usually bad SQL/code or an index), and re-measure.
What is the practical takeaway on Performance tuning?
Database: expensive SQL, missing indexes, outdated statistics (usually the first bottleneck). Custom ABAP: SELECT-in-loop, SELECT *, nested loops on big tables. Buffers/memory: under-sized program/table buffers.
What tends to go wrong with Performance tuning?
Tuning without measuring. Ignoring custom-code SQL. Under-sized DB/buffers.