master
md 110 lines 4.78 KB
Rendered Raw
1 # Configure Netdata Parents to centralize metrics
2
3 ## How Metrics Centralization Works
4
5 ```mermaid
6 flowchart BT
7 C1("**Netdata Child 1**<br/><br/>Production System")
8 C2("**Netdata Child 2**<br/><br/>Production System")
9 C3("**Netdata Child N**<br/><br/>Production System")
10 P1("**Netdata Parent 1**<br/><br/>Centralization Point")
11 C1 -->|stream| P1
12 C2 -->|stream| P1
13 C3 -->|stream| P1
14
15 %% Style definitions
16 classDef child fill:#ffeb3b,stroke:#000000,stroke-width:3px,color:#000000,font-size:16px
17 classDef parent fill:#4caf50,stroke:#000000,stroke-width:3px,color:#000000,font-size:16px
18
19 %% Apply styles
20 class C1,C2,C3 child
21 class P1 parent
22 ```
23
24 ## Installation
25
26 A **Netdata Parent** is a regular Netdata Agent configured to receive streaming metrics from Netdata Children (also normal Agents!).
27
28 To install a Netdata Parent:
29
30 1. Install Netdata on your Parent node.
31
32 For installation instructions, see the [Installation section](/packaging/installer/README.md).
33
34 2. After installation, configure `stream.conf` to enable reception from Children.
35
36 See [Configuring Metrics Centralization Points](/docs/observability-centralization-points/metrics-centralization-points/configuration.md) for detailed configuration instructions.
37
38 ## Key Features
39
40 :::important
41
42 **Netdata Streaming and Replication**
43
44 Copies **recent past samples** (replication) and **real-time new samples** (streaming) from production systems (**Netdata Children**) to **metrics centralization points** (**Netdata Parents**). **Netdata Parents** store the database for these metrics based on **retention settings**.
45
46 :::
47
48 ### How Your Systems Connect
49
50 <details>
51 <summary><strong>Netdata Child Behavior</strong></summary><br/>
52
53 - Each **Netdata Child** can stream to **only one** Netdata Parent at a time.
54 - Multiple **Netdata Parents** can be configured for **high availability**, but only the **first working one** will be used.
55
56 <br/>
57 </details>
58
59 <details>
60 <summary><strong>Netdata Parent Capabilities</strong></summary><br/>
61
62 - Receives metric samples **from multiple Netdata Children**.
63 - Can **re-stream** received metrics to another **Netdata Parent**, forming an **infinite hierarchy** of Parents.
64 - Supports **Netdata Parents Clusters** for **high availability**.
65
66 <br/>
67 </details>
68
69 ## Feature Comparison
70
71 | Feature | Netdata Child (Production System) | Netdata Parent (Centralization Point) |
72 |----------------------------|---------------------------------------------------|-------------------------------------------------------|
73 | **Metrics Retention** | Minimal retention; can use `ram` or `alloc` mode. | Stores metrics for all connected systems. |
74 | **Machine Learning** | Can be disabled (default: enabled). | Runs anomaly detection for all connected systems. |
75 | **Alerts & Notifications** | Can be disabled (default: enabled). | Monitors health and sends alerts for all systems. |
76 | **API & Dashboard** | Can be disabled (default: enabled). | Hosts the dashboard using its own retention settings. |
77 | **Exporting Metrics** | Optional (default: enabled). | Exports all collected metrics. |
78 | **Netdata Functions** | Child must be online to function. | Forwards function requests to connected Children. |
79 | **Netdata Cloud** | Not required. | Registers all connected systems to Netdata Cloud. |
80
81 ## **Supported Configurations**
82
83 ### **For Netdata Children**
84
85 - **Full Mode (Default)**:
86 - All Netdata features are enabled (machine learning, alerts, notifications, dashboard, etc.).
87 - **Thin Mode**:
88 - Only collects and forwards metrics to a Parent.
89 - Some local retention is kept to handle network issues, but all other features are disabled.
90
91 ### **For Netdata Parents**
92
93 - **Standalone**:
94 - A single Parent in the infrastructure or the top-most Parent in a hierarchy.
95 - **Cluster**:
96 - A group of Parents that share the same data from the same Children.
97 - Provides **high availability**.
98 - **Proxy**:
99 - Stores received metrics locally and **forwards them** to a higher-level Parent (Grand Parent).
100
101 ### **Cluster Configuration**
102
103 - A Cluster consists of **circular Proxy nodes**, where each Parent acts as a Parent to the others.
104 - Only the **top level** of a multi-level hierarchy can be configured as a cluster.
105
106 ## **Best Practices**
107
108 For detailed guidelines, check [Best Practices for Observability Centralization Points](/docs/observability-centralization-points/best-practices.md).
109
110 To understand how node connection states (Live, Stale, Offline) work with Parent-Child setups, see [Node States and Transitions](/docs/netdata-cloud/node-states-and-transitions.md).