perf: thread local stat caching
Abierto
area/perf
help wanted
- Lenguaje dominante
- C++
- Estrellas
- 28.9k
- Forks
- 5.6k
- Merge medio
- 1 d 22 h
- PR fusionados (30 d)
- 430
Descripción
Running Envoy on very high core count boxes is going to lead to atomic contention for counters and gauges. We should implement the option of doing thread local caching of the stats, and then flushing them to the central global representation.
Guía de contribución
Evaluación
Este issue todavía no se ha evaluado.