The Top 8 Reasons for Your Transaction Performance Problems
summary
Lists eight concrete causes and fixes rather than vague examples: misconfigured max server memory and optimize for ad hoc workloads settings that blow plan cache, disk latency over 5–10 ms from shared storage or 300‑VM oversubscription, synchronous transaction log growth and 60 KB log‑buffer flush behavior, page splits from poor clustered indexes, excessive nonclustered indexes, page latch hotspots, CXPACKET from bad parallel plans (raise cost threshold to ~40–50), and a claim that ~70% of blocking traces back to resources not code.
AI-written summary · about
chapters (17)
- 0:00Start
- 5:00Agenda
- 5:53About me
- 6:27The history behind the session
- 9:22The query travel trough SQL Server
- 10:19A modification travel trough SQL Server
- 11:44Possible pitfalls of data modification's performance problems
- 13:40Critical Memory configuration settings
- 15:00What does the overwhelmed mean?
- 17:34Poor physical layout of the database files
- 19:37"Do not play with tempting settings" rules
- 21:17Flushing log records
- 24:25Page Splits
- 26:17Over-indexing
- 27:56Contention latches (page concurrency)
- 32:39Lock Blocking
- 35:29Summary
related talks
47mArchitecting for High Performance SQL Server on Virtual Machines
57mYour Performance Tuning Check List
31m10 Costly Database Performance Mistakes (and How to Fix Them)
54mYour Performance Tuning Check List
43mPerformance Tuning, Getting the Biggest Bang for Your Buck
56mQuery Planning Gone Wrong
58mTroubleshooting CPU Performance Issue for SQL Developers
17mTop 10 transactional database mistakes, and how to avoid them