---
title: "Discussion on \"NGINX Worker Processes and Connections\" | Hashnode"
description: "Discussion on \"NGINX Worker Processes and Connections\". Worker processes\nWhen we start the NGINX service, we get a master process which we can view using systemctl. This master process then spawns worker processes that are used to serve requests. The default is one worker process.\nWe can change the no. of..."
url: "https://hashnode.com/posts/nginx-worker-processes-and-connections/64ba1b91d5b4c9b086939dce"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [NGINX Worker Processes and Connections](https://blog.melvincv.com/nginx-worker-processes-and-connections?utm_source=hashnode&utm_medium=feed)

Worker processes When we start the NGINX service, we get a master process which we can view using systemctl. This master process then spawns worker processes that are used to serve requests. The default is one worker process. We can change the no. of...

Jul 21, 2023 · 2 min read · blog.melvincv.com  
Read the article

-   [#nginx](https://hashnode.com/tag/nginx)
-   [#nginx-configuration-guide](https://hashnode.com/tag/nginx-configuration-guide)
-   [#worker](https://hashnode.com/tag/worker)
-   [#connection](https://hashnode.com/tag/connection)

## Written by

[Melvin C Varghese (melvincv)](https://hashnode.com/@melvincv)

@melvincv

  
[View profile](https://hashnode.com/@melvincv)

DevOps Engineer

I am a DevOps Consultant. 3 years experience with Cloud technologies. 9 years total IT experience in the Linux, Networking and Data Visualization domains. Love listening to music and playing older PC games.

## Responses 1

Markdown supported · ⌘↵ to post

[Edge Ops](https://hashnode.com/@edgecacheops)

[Oct 31](https://hashnode.com/posts/nginx-worker-processes-and-connections/64ba1b91d5b4c9b086939dce/comment/6ae585363a0cd7a5072d66cc)

Clear explanation of master vs workers and why worker\_connections is not just "max RPS." In reverse-proxy edge roles I size workers to CPU cores (auto is fine) and set worker\_connections for concurrent client sockets plus upstream sockets, because each proxied request can hold both sides open under keepalive. Underestimating that number shows up as accept queue growth and intermittent 502s long before CPU saturates.

Ops check after a tune: watch active and waiting connections under load, confirm worker\_connections \* worker\_processes leaves headroom above peak, and keep multi\_accept on only when the accept queue actually builds. If waiting climbs while CPU is idle, raise connections or shorten idle keepalive timeouts before adding hardware. Log worker exits and reopen cycles after reload — a sudden drop in worker count after config change usually means a limit was hit, not a traffic cliff.
