Add article for hypernode-haproxyctl

docs/hypernode-platform/cluster.md

@@ -85,6 +85,8 @@ least connections.
 Web traffic is terminated by NGINX, which (optionally) forwards the request to varnish, which can return a cached page or send the
 request to [HAProxy](https://www.haproxy.org/), which sends the request to the application servers.
 
+To manage the HAProxy service, you can use our utility command `hypernode-haproxyctl`. Learn more about this command in our [hypernode-haproxyctl documentation](tools/how-to-use-hypernode-haproxyctl-tool.md).
+
 ### Managing NGINX
 
 To add custom NGINX rules for handling specific things on the loadbalancer (like serving assets), the `loadbalancer.` prefix can be

docs/hypernode-platform/tools/how-to-use-hypernode-haproxyctl-tool.md

@@ -0,0 +1,51 @@
+---
+myst:
+  html_meta:
+    description: 'With the hypernode-haproxyctl tool you can inspect and manage the
+      HAProxy service. Learn how to use the hypernode-haproxy CLI Tool. '
+    title: How to use the hypernode-haproxyctl CLI Tool?
+---
+
+# How to Use the hypernode-haproxyctl CLI Tool
+
+Hypernode clusters use [HAProxy](https://www.haproxy.org/) to distribute HTTP requests from the loadbalancer to the application servers. With the `hypernode-haproxyctl` CLI tool you can inspect and manage the HAProxy service.
+
+## Viewing the current status
+
+With the `status` subcommand, you can see the current status of HAProxy and its backends.
+
+```console
+app@abcdef-examplelb-magweb-tbbm ~ $ hypernode-haproxyctl status
+HAProxy v2.6.12-1+deb12u3 (released: 2025/10/03, uptime: 98d 22h33m56s)
+FRONTEND (OPEN): request rate(160), session rate(7), conn rate(7)
+exampleapp1 (UP): request rate(0), session rate(25), conn rate(0)
+exampleapp2 (UP): request rate(0), session rate(32), conn rate(0)
+exampleapp3 (UP): request rate(0), session rate(28), conn rate(0)
+exampleapp4 (UP): request rate(0), session rate(38), conn rate(0)
+exampleapp5 (UP): request rate(0), session rate(35), conn rate(0)
+BACKEND (UP): request rate(0), session rate(160), conn rate(0)
+```
+
+The output has three sections:
+
+- **FRONTEND**: The internet-facing entry point. Shows how many requests and connections are coming in per second.
+- **exampleapp1-N**: Your individual app servers. Common states are `UP`, `DOWN` and `DRAIN`.
+- **BACKEND**: The combined pool of all app servers. `UP` means at least one server is healthy.
+
+If a server shows `DOWN` unexpectedly, it likely needs attention. `DRAIN` and `MAINTENANCE` are usually intentional states used during maintenance.
+
+## Managing backends
+
+When troubleshooting or performing maintenance on an app server, you can use the `drain` subcommand to gracefully remove it from rotation:
+
+```console
+app@abcdef-examplelb-magweb-tbbm ~ $ hypernode-haproxyctl drain exampleapp1
+```
+
+The `drain` command allows existing connections to finish while preventing new traffic from being sent to the server. This is ideal for performing maintenance without disrupting active users.
+
+Once maintenance is complete, you can use the `enable` subcommand to bring the server back into rotation:
+
+```console
+app@abcdef-examplelb-magweb-tbbm ~ $ hypernode-haproxyctl enable exampleapp1
+```