> ## Documentation Index
> Fetch the complete documentation index at: https://metoro.io/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Overview

> Understanding the Metoro architecture and components

This deployment brings Metoro directly into your infrastructure.

While we strive to make installation and maintenance as boring as possible, it often helps to know how Metoro is architected so you can understand what you're running and how to operate it.

## Architecture Overview

Metoro consists of two main parts:

1. **Metoro Hub**

   The central control plane that processes, stores, and serves observability data. In the cloud installation, this is managed by Metoro, but here you deploy and manage it yourself.

2. **Metoro Agents**

   Lightweight agents deployed on each Kubernetes cluster that collect and forward data to the hub. These run the same way as the cloud installation.

```mermaid theme={null}
graph TB
    subgraph "Your Kubernetes Clusters"
        subgraph "Cluster 1"
            NA1[Node Agents<br/>DaemonSet]
            E1[Exporters]
            P1[Your Workloads]
            NA1 -.->|eBPF| P1
            NA1 -->|Telemetry| E1
        end
        subgraph "Cluster 2"
            NA2[Node Agents<br/>DaemonSet]
            E2[Exporters]
            P2[Your Workloads]
            NA2 -.->|eBPF| P2
            NA2 -->|Telemetry| E2
        end
    end

    subgraph "Metoro Hub"
        I[Ingester]
        API[API Server]
        CH[(ClickHouse)]
        PG[(PostgreSQL)]
        T[Temporal]

        I --> CH
        API --> CH
        API --> PG
        API --> T
        T --> PG
    end

    E1 -->|OTLP| I
    E2 -->|OTLP| I

    U[Users] -->|Web UI/API| API
```

We recommend that Metoro is deployed in a hub and spoke architecture. One organization wide instance of the Metoro Hub into which all telemetry feeds.
This has a number of benefits:

1. Comparison of data and services across clusters is possible
2. Operationally less overhead, one installation of stateful components is easier to handle than N.
3. Resource efficient, each Metoro installation comes with fixed overhead, centralizing it leads to less waste and more efficient resource usage.

<CardGroup cols={2}>
  <Card title="Metoro Hub" icon="server" href="/docs/on-premises/10.x/architecture/metoro-hub/overview">
    The self-hosted control plane, data stores, UI/API, ingestion, and workflow components.
  </Card>

  <Card title="Metoro Exporter" icon="satellite-dish" href="/docs/on-premises/10.x/architecture/metoro-exporter/overview">
    The lightweight per-cluster components that collect telemetry and send it to the hub.
  </Card>
</CardGroup>