How Frappe v14 is blAziNgLy fast
summary
v14 cuts response times, memory use, bundle size and DB footprint with concrete fixes: a buggy foreign‑key cache causes hundreds of duplicate queries and is fixed using frappe recorder and cProfile; over‑caching of documents inflates memory and is corrected with memory_profiler yielding near‑constant memory; sampling with py-spy removes boilerplate overhead and drops typical requests ~50%; swapping the MySQL client promises 2.5× performance for queries; bundle dedupe trims JS from ~5MB to ~4MB and DB cleanup halves disk from ~5GB to ~2.4GB.
AI-written summary · about
chapters (6)
- 0:00Start
- 3:09Known Problems in Performance
- 5:09Recorder
- 7:45Memory Profiler
- 19:16Skeleton Loaders
- 22:25Takeaways
related talks
34mNeed for Speed: Removing speed bumps in your APIs
45mCutting latency in half: What actually worked—and what didn’t
31mNo more slow requests
37mPerformant Django
39mWhat is Performance Driven Development and why aren't you doing it?
01h 15mSpeedupRails, Speedup Your Code
01h 02mPerformance
01h 02mSpeeding up the Web with PHP 7