docs: correct Hetzner hardware and pricing in cluster design and buying guide

Author: btwshivam

docs/cluster-design.md

@@ -477,25 +477,27 @@ kubectl get pods --all-namespaces -o json | jq -r '
 | **Network** | 10G NIC available | Fast Ceph replication |
 | **Cost** | Higher | Better for production |
 
-#### Recommended Hardware: AX52
+#### Recommended Hardware: AX102
 
 | Component | Specification | Purpose |
 | ----------- | --------------- | --------- |
-| **CPU** | AMD Ryzen 9 5950X (16c/32t) | High-performance workloads |
-| **RAM** | 128 GB DDR4 ECC | Large memory for apps + Ceph |
+| **CPU** | AMD Ryzen 9 7950X3D (16c/32t) | High-performance workloads |
+| **RAM** | 128 GB DDR5 ECC | Large memory for apps + Ceph |
 | **Storage** | 2× 3.84 TB NVMe | Ceph OSD + Local PV |
 | **Network** | **10 Gbps NIC** | Fast Ceph replication, low latency |
 
+See the [Hetzner purchasing guide](./guides/hetzner-server-buy-guides.md) for sizing and current prices.
+
 ```mermaid
 graph TB
     subgraph cluster["🏭 Production Cluster on Dedicated Servers"]
-        subgraph cp["🔐 Control Plane (3× CAX41 ARM64)"]
+        subgraph cp["🔐 Control Plane (3× CAX11 ARM64)"]
             cp1["CP-1<br/>Ampere"]
             cp2["CP-2<br/>Ampere"]
             cp3["CP-3<br/>Ampere"]
         end
         
-        subgraph workers["⚙️ Workers (3× AX52 + 10G NIC)"]
+        subgraph workers["⚙️ Workers (3× AX102 + 10G NIC)"]
             w1["Worker-1<br/>Ceph OSD<br/>Apps"]
             w2["Worker-2<br/>Ceph OSD<br/>Apps"]
             w3["Worker-3<br/>Ceph OSD<br/>Apps"]
@@ -627,14 +629,14 @@ databases:
 # Cluster configuration
 controlPlane:
   count: 3
-  hardware: AX41-NVMe          # Dedicated servers
+  hardware: AX102              # Dedicated servers
   taints:
     - key: node-role.kubernetes.io/control-plane
       effect: NoSchedule
 
 workers:
   count: 3+
-  hardware: AX41               # With Ceph OSD disks
+  hardware: AX102              # With Ceph OSD disks
 
 storage:
   primary: rook-ceph-block     # Ceph for all persistent storage
@@ -669,7 +671,7 @@ kubectl taint nodes <cp-node> node-role.kubernetes.io/control-plane:NoSchedule
 2. Add more worker nodes
 3. Migrate to Ceph storage
 
-See: [Hetzner CSI Volume Limit Debugging](../guides/storage/hetzner/hetzner-csi-volume-limit-debugging.md)
+See: [Hetzner CSI Volume Limit Debugging](https://gitea.obmondo.com/EnableIT/wiki/src/branch/master/guides/storage/hetzner/hetzner-csi-volume-limit-debugging.md)
 
 ### "Database not recovering after node failure"
 

docs/guides/hetzner-server-buy-guides.md

@@ -1,4 +1,3 @@
-
 # Hetzner Bare Metal & HCloud Server Purchasing Guide
 
 This document provides guidance for purchasing Hetzner servers for Kubernetes deployments, including recommended hardware,
@@ -26,7 +25,6 @@ For **bare metal clusters**, we recommend:
 
 * **Single server setup**: One node (suitable for testing or small deployments).
 * **High-availability setup**: 3 control plane nodes + 3 worker nodes for redundancy.
-(we recommend running the control plane on VMs to save on resources and cost).
 
 ### Best Practices
 
@@ -67,19 +65,35 @@ For **bare metal clusters**, we recommend:
 
 ## 3. HCloud Server Recommendations
 
-* **Control Plane Node**: CAX11 (2 vCPU, 4 GB RAM, 40 GB storage, ~4€/month)
-* **Worker Node**: CAX41 (16 vCPU, 32 GB RAM, 320 GB storage, ~23€/month)
+* **Control Plane Node**: CAX11 (2 vCPU, 4 GB RAM, 40 GB storage, 5.99€/month)
+* **Worker Node**: CAX41 (16 vCPU, 32 GB RAM, 320 GB storage, 40.99€/month)
+
+Add 0.50€/month per node for the IPv4 address. Prices are list, as of the
+[15 June 2026 adjustment](https://docs.hetzner.com/general/infrastructure-and-availability/price-adjustment/).
+Hetzner changed prices three times in the first half of 2026, so re-check that page before you quote a total.
 
 ### Important Notes
 
 * HCloud nodes have a **limit of 16 volumes per node**.
 * Volume attachment is subject to **location limitations per country**, so plan accordingly.
 
+### Outbound Traffic
+
+Traffic is usually the largest hidden cost difference against a hyperscaler, so budget it explicitly:
+
+* HCloud servers in **Germany and Finland include 20 TB outbound per server per month**, pooled across the project.
+* **US locations include roughly 1 TB**, and **Singapore as little as 0.5 TB**, so the same cluster costs
+  materially more outside the EU.
+* Overage is **€1/TB** in the EU locations and **€7.40/TB** in Singapore.
+* The allowance is fair use, not a contract. It pools across the project rather than guaranteeing headroom
+  to one hot node.
+* Bare metal servers are sold with unlimited traffic on the standard 1 Gbit uplink.
+
 ---
 
-## 4. Hybrid Cluster  (Recommended setup)
+## 4. Hybrid Cluster (Recommended setup)
 
-* **Control Plane**: CAX11 HCloud server (~4€/month)
+* **Control Plane**: CAX11 HCloud server (5.99€/month, plus 0.50€ for the IPv4)
 * **Worker Node**: Bare Metal
 
   * RAM: 4 × 16384 MB DDR4
@@ -96,7 +110,7 @@ For **bare metal clusters**, we recommend:
 | Layout            | Nodes                                               | Storage       | RAM           | Notes                               |
 | ----------------- | --------------------------------------------------- | ------------- | ------------- | ----------------------------------- |
 | Single Bare Metal | 1 node                                              | 2 × 1 TB NVMe | 4 × 16 GB     | Small/test cluster                  |
-| HA Bare Metal     | 3 control + 3 worker                                | 2 × 10 TB HDD | 4 × 16 GB     | Production with redundancy          |
+| HA Bare Metal     | 3 control + 3 worker                                | 2 × 1 TB NVMe | 4 × 16 GB     | Production with redundancy          |
 | All HCloud        | 3 control (CAX11) + 3 worker (CAX41)                | CAX41 default | CAX41 default | Cloud cluster with scaling          |
 | Hybrid            | 1 or 3 control (CAX11) + 1 or 3 worker (bare metal) | 2 × 1 TB NVMe | 4 × 16 GB     | Internal high-performance workloads |