TopicsPerformance
Performance
Profiling, flame graphs, latency budgets, and load tests you can defend in interviews.
Common tags: profiling, flame-graphs, latency, p99
- Performance
Performance Engineering — Profiling, Flame Graphs, Latency Budgets & Hot Paths
Cluster · Performance Engineering
Measure→hot-path→fix→re-measure loop; latency vs throughput vs capacity; when to profile CPU vs memory vs wait/IO.
Open study →- performance
- profiling
- flame-graphs
- latency
- p99
- hot-paths
- interview
- Performance
Microbenchmarks vs Load Tests — Lies, Variance & Capacity
Cluster · Performance Engineering
DCE/warmup/noise lies in microbench; open vs closed load; coordinated omission; capacity vs regression.
Open study →- performance
- microbenchmarks
- load-testing
- k6
- wrk
- variance
- capacity
- interview
- Performance
Memory Profiling — Allocations, Leaks & GC Pressure
Cluster · Performance Engineering
Allocation rate vs retained heap; leaks vs caches; GC pressure on p99.
Open study →- performance
- memory
- allocations
- leaks
- gc
- tracemalloc
- heap
- interview
- Performance
Latency Budgets & Critical Path — Where the p99 Lives
Cluster · Performance Engineering
Budget each hop; find the critical path; fan-out p99 ≈ max of children; averages lie.
Open study →- performance
- latency
- p99
- critical-path
- budgets
- fan-out
- interview
- Performance
Flame Graphs & Stack Collapse — Reading Hot Paths
Cluster · Performance Engineering
Width = sample share, not one slow call; how to read plateaus and avoid common misreads.
Open study →- performance
- flame-graphs
- stack-collapse
- hot-paths
- brendan-gregg
- interview
- Performance
CPU & Wall-Clock Profiling — Sampling vs Instrumentation
Cluster · Performance Engineering
Sampling vs instrumentation; on-CPU vs wall-clock; GIL/event-loop traps that make “CPU” misleading.
Open study →- performance
- profiling
- sampling
- instrumentation
- cpu
- wall-clock
- py-spy
- pprof
- interview