How console.log Silently Kills Your Server Performance in Production

Open-source developer Pooya Parsa highlighted a significant performance cost hidden in routine server logging, sparking discussion among developers including Vercel founder Guillermo Rauch. Benchmarks from the srvx request logger showed that writing logs to /dev/null — with no disk or network overhead — still reduced throughput by around 21%, while logging to a terminal pushed that loss to 34%. The culprit was repeated locale-aware timestamp generation and per-request stream writes on every incoming request. Replacing these with cached timestamps and batched writes recovered 60–85% of the lost performance. The trade-off, as Rauch noted, is that buffered logs risk being lost if a process crashes before the buffer is flushed, making structured production logging tools a safer and smarter choice than console.log.