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

# Chronosphere Telemetry Pipeline architecture

Chronosphere Telemetry Pipeline uses different components to create, modify, and
run pipelines.

```mermaid actions={false} theme={null}
    flowchart TB
  accTitle: Telemetry Pipeline architecture diagram
  accDescr: An diagram of Telemetry Pipeline architecture.
  subgraph cluster ["Kubernetes or K3s cluster"]
    direction TB
    subgraph namespace [Namespace]
        direction TB
    operator(Core Operator)
    CRDs@{shape: lined-cylinder, label: "Pipeline<br>custom<br>resources"}
    instance(Core Instance)
    operator-."watches"-.->CRDs
    instance--"creates and<br>updates"---->CRDs
        subgraph daemonset ["Pipeline (DaemonSet)"]
            pod3([Pod])
        end
        subgraph statefulset ["Pipeline (StatefulSet)"]
            pod4([Pod])
            pod5([Pod])
        end
        subgraph deployment ["Pipeline (Deployment)"]
            pod1([Pod])
            pod2([Pod])
        end
    operator--"creates and <br>modifies"--->deployment
    operator--"creates and <br>modifies"--->statefulset
    operator--"creates and <br>modifies"--->daemonset
    end
  end
    subgraph chronosphere [Chronosphere]
        backend[Telemetry Pipeline<br>backend]
        frontend[Telemetry Pipeline<br>web interface]
        frontend --"modifies"-->backend
    end
  instance<-."syncs<br>status"-..->backend
  instance--"fetches<br>config"--->backend
  backend ~~~ CLI((Pipeline CLI)) --"modifies"--> backend
  CLI ~~~ backend
  CLI --"modifies"--> instance

style namespace stroke:#5a78c4,stroke-dasharray: 5 5

classDef greenFill stroke:#60E684,fill:#2FBF71,stroke-width:2px,color:#FFFFFF
class pod1,pod2,pod3,pod4,pod5,frontend,backend greenFill;

classDef greenLine stroke:#60E684,stroke-width:2px
class pod1,pod2,pod3,pod4,pod5 greenLine;

classDef blueFill fill:#5a78c4,stroke-width:2px,color:#FFFFFF
class daemonset,deployment,statefulset,CLI blueFill;

classDef blueLine stroke:#5a78c4,stroke-width:2px
class cluster,daemonset,deployment,statefulset,chronosphere,CLI blueLine;
```

Telemetry Pipeline is built on Kubernetes. You can either
[install Telemetry Pipeline](/ingest/pipeline/v2/install) in an existing Kubernetes cluster
or in a Linux environment. If you install Telemetry Pipeline in a Linux environment,
the installer automatically deploys a K3s cluster, then installs Telemetry Pipeline
components in that cluster.

Each Telemetry Pipeline installation includes a Core Operator, which is a
[Kubernetes operator](https://kubernetes.io/docs/concepts/extend-kubernetes/operator/)
that creates and modifies pipelines. A different component called a Core Instance
creates and updates the custom resources that the Core Operator uses when it creates
or modifies pipelines. The Core Instance also syncs the status of your cluster
with the Telemetry Pipeline backend.

Each pipeline is a workload that runs within your cluster and contains one or more Pods.
By default, pipelines run as
[Deployments](https://kubernetes.io/docs/concepts/workloads/controllers/deployment/),
but other [workload types](/ingest/pipeline/v2/configure/kubernetes/workloads) are supported as well.

You can use [Pipeline CLI](/ingest/pipeline/pipeline-cli) and the
[Telemetry Pipeline web interface](https://core.calyptia.com/) to configure your
Telemetry Pipeline installation and to create and modify pipelines.