# HAProxy Cluster

**URL:** https://discourse.haproxy.org/t/haproxy-cluster/5365
**Category:** Help!
**Created:** [June 21, 2020, 8:48pm UTC](https://discourse.haproxy.org/t/haproxy-cluster/5365 "2020-06-21T20:48:52Z")
**Posts on this page:** 3
**Page:** 1

<div class="post-metadata">

### Author: ![hac3ru](https://avatars.discourse-cdn.com/v4/letter/h/ee59a6/32.png) [@hac3ru](https://discourse.haproxy.org/u/hac3ru)
#### Post date: [June 21, 2020, 8:48pm UTC](https://discourse.haproxy.org/t/haproxy-cluster/5365/1 "2020-06-21T20:48:52Z")

</div>

Hello,

I have got a large number of HAProxies for a single app (over 20 servers). The issue I’m facing is that, even though I have set max number of conns per backend server, since the haproxies are unaware of other haproxies, the load on the backend machines is far from balanced. Is there any way to create a “true” cluster of haproxies, where each proxy knows of all the other proxies and knows how many each backend server has established, how many more it can process and so on, based on some rules?

I hope my question is clear…

Thank you.

---

<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: [June 22, 2020, 6:19am UTC](https://discourse.haproxy.org/t/haproxy-cluster/5365/2 "2020-06-22T06:19:09Z")

</div>

No, this is not possible and probably never will be. Global rate-limiting (or maxconn based queueing) with synchronization doesn’t work, because it’s performance critical data that cannot be synchronized in time.

You need to design around that.

---

<div class="post-metadata">

### Author: ![hac3ru](https://avatars.discourse-cdn.com/v4/letter/h/ee59a6/32.png) [@hac3ru](https://discourse.haproxy.org/u/hac3ru)
#### Post date: [June 22, 2020, 6:34am UTC](https://discourse.haproxy.org/t/haproxy-cluster/5365/3 "2020-06-22T06:34:51Z")

</div>

Thank you.
