🧵I took my from-scratch Redis server (in Java) from 1,200 ops/s → 83,000 ops/s (+70x) using flame graphs. No magic. Just profiling and fixing the real bottlenecks.
Here's exactly what I found:
the go memory model reading - done
lecture 5: go threads, raft - done
- the lecture was about some debugging techniques and concurrency patterns
- some common bugs while implementing raft
next: raft paper.
wish me luck