# Non-global maxconn when nbproc \> 1

**URL:** <https://discourse.haproxy.org/t/non-global-maxconn-when-nbproc-1/2965>\
**Category:** Help!\
**Created:** [September 6, 2018, 8:15pm UTC](https://discourse.haproxy.org/t/non-global-maxconn-when-nbproc-1/2965 "2018-09-06T20:15:07Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![cvcrckt](https://avatars.discourse-cdn.com/v4/letter/c/d07c76/32.png) [@cvcrckt](https://discourse.haproxy.org/u/cvcrckt)\
**Post date:** [September 6, 2018, 8:15pm UTC](https://discourse.haproxy.org/t/non-global-maxconn-when-nbproc-1/2965/1 "2018-09-06T20:15:07Z")

</div>

I understand that the global maxconn is the maximum number of concurrent connections _per process_. If I want my HAProxy to be able to handle 120k concurrent connections max and I have specified ‘nbproc 4’, I will set the global maxconn to 30000.

Imagine I have a single frontend and a single backend defined, and I want to allow for 120k max concurrent connections to them, too. When I am running with ‘nbproc 4’, should maxconn for the frontend binds and for the backend also be set to 30000? Or is maxconn in these sections specifying _aggregate_ maximum connections for the frontend/backend across all running haproxy processes?

If it makes a difference, I am using HAProxy 1.5.

---

<div class="post-metadata">

**Author:** ![lukastribus](https://avatars.discourse-cdn.com/v4/letter/l/7ea924/32.png) [@lukastribus](https://discourse.haproxy.org/u/lukastribus)\
**Post date:** [September 6, 2018, 8:40pm UTC](https://discourse.haproxy.org/t/non-global-maxconn-when-nbproc-1/2965/2 "2018-09-06T20:40:40Z")

</div>

Everything you configure in the configuration is for a single process, there are no inter-process aggregates when talking about maxconn.

Backends actually don’t use maxconn at all; rather you should be setting per-server maxconn values (to avoid overloading single servers).

So global, frontend and backend-server (but not the backend itself) maxconn values are what is important.