Performance investigations

Case studies (aka performance crimes)

1. Concurrent collection misuse: how the slowness could remain undetected for ages
2. Careless allocations: don’t trust developers saying allocations are cheap
3. Wrong size detection: how come server is out of memory when configured not to
4. No respect for mainstream flow: performance win by optimizing for single scenario
5. Configuration to kill performance: one setting can halt application speed
6. Memory leak investigation: how to check why server is out of RAM
7. Outdated reports in Sitecore Analytics: you could not even guess why
8. Price for non-used features: are you sure disabling feature is all it takes?
9. How much faster can software be?: is Wirth’s law reality or not?
10. Second life for Fast://: can the powerful technology get a second life?
11. Database optimization – reading same data 7 times faster
12. Hidden culture-specific parsing – spending an extra second out of blue
13. Params – why all performance-critical code has overloads without it?
14. Reckless interlocked operations – why could that be painful?
15. Concat – what about joining strings together?
16. Reflection – how costly could it be?
17. Non-unique strings – consume less RAM without no compromises in functionality

Docker

• How to measure performance in containers?
• Is docker file system slow?
• Can configuration be trusted?

Development practices

• Unit testing in Sitecore – why Sitecore is not testable cliche is obsolete
• Make unit tests public – what benefits could be introduced by exposing unit tests?

Sitecore Analytics

• Polluted reports – over 20% of data you collect could be from robots
• How often the site is visited from specific place?