Envoy vs Nginx at scale: why Dropbox and Cloudflare moved off Nginx
Published 2026-10-19 · Built only from first-party engineering write-ups in our database. Every number below comes from the linked source.
Dropbox is the direct case: it moved its edge traffic proxy from Nginx to Envoy for better observability and extensibility. Cloudflare also moved off Nginx, but to Pingora, its own Rust proxy that reuses connections better and needs less CPU and memory, at 1 trillion requests a day. OpenAI uses Envoy for connection fan-in in its Habitat storage platform.
Nginx vs Envoy: side by side
| Nginx | Envoy | |
|---|---|---|
| Stories in our database | 2 | 2 |
| Companies | Cloudflare and Dropbox | Dropbox and OpenAI |
| Published | 2022 (1 undated) | 2026 (1 undated) |
| Techniques alongside it | connection reuse (1), load balancing (1), proxy migration (1), rewrite (1) | load balancing (2), change data capture (1), connection pooling (1), proxy migration (1) |
| Scale numbers stated in sources | 1 trillion requests/day (Cloudflare) | 70 million+ requests/sec (OpenAI) |
| Direct comparison story | Dropbox: How we migrated Dropbox from Nginx to Envoy; Cloudflare: How we built Pingora, the proxy that connects Cloudflare to the Internet | (same story) |
Verdict
Both teams that left Nginx did it for reasons specific to very large fleets: observability and extensibility at Dropbox, connection reuse and efficiency at Cloudflare. These are production stories, not benchmarks.
Counts are stories in How They Scaled tagged with each side, see the Nginx and Envoy pages. They describe what these teams wrote about, not market share or benchmark results.