Why databases cry at night?
summary
Dives into concrete NoSQL failure modes and fixes: shows how a select N+1 anti-pattern becomes dozens of queries but embedding order details drops that to two queries; demonstrates a 30-join query cut to three joins and runtime falling from ~4 minutes to ~300 ms; profiles JSON serialization 30–40% CPU with json.net and switches to an offset-based format to avoid allocations; reveals production surprises like 200 DBs I/O storm at startup and 300 background indexes exhausting threads — emphasizes query/index planning, monitoring, and picking optimistic vs pessimistic concurrency.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 12:49MongoDB example
- 16:59Query complexity
- 21:32Reducing joins
- 22:59The root cause
- 24:02Another riddle
- 31:04Resources
- 37:10Concurrency
related talks
58mWhy Databases Cry At Night
45mNoSQL: Death to Relational Databases
41mUsing Databases to Pull Your Application Weight
35mAn Ode to 17 Databases in 39 minutes
18mFrom SQL to NOSQL - a gentle intro for Devs
59mWhy Databases Cry At Night
54mSay Yes to NoSQL for the .NET SQL Developer
53mFrom RDBMS to NoSQL (PRT314)