Networking

Use the information on this page to configure different networking settings in Chronosphere Telemetry Pipeline.

Ports

Pipelines use various ports (also referred to as endpoints) to send and receive data. These ports correspond to different components of each pipeline.

View pipeline ports

To view a list of open ports for a pipeline, use the following command:

calyptia get endpoints --pipeline PIPELINE

Replace PIPELINE with the name of your pipeline.

Each item in the resulting list includes a protocol type, a frontend port, and a backend port.

Plugins

Most source plugins and some destination plugins expose an associated port. Plugins that expose a port will set a default value for that port, but you can override this value when you configure plugins through the Telemetry Pipeline web interface or through pipeline configuration files.

Health check pipelines

By default, the health check pipeline within each Core Instance exposes an HTTP server on TCP port 2020.

Frontend and backend ports

Each pipeline component uses a triplet that consists of a frontend port, a backend port, and a protocol type. Frontend ports are exposed externally and let pipelines communicate with different services to send and receive telemetry data. Backend ports are only used within a pipeline’s internal Kubernetes infrastructure to forward traffic to pods.

When you set a port number for a pipeline component, the same port number is applied to both the frontend and backend ports for that component. These frontend and backend port numbers stay in sync unless a conflict triggers automatic reassignment for the frontend port.

Uniqueness

Within a single pipeline, each backend port must use a unique port number and protocol. Within a Core Instance, backend ports can overlap across pipelines, but each frontend port in the Core Instance must use a unique port number and protocol. However, different components or pipelines can share a port number if one port uses TCP and the other uses UDP.

For example, two pipelines in the same Core Instance can’t both have plugins that expose TCP port 1600, but Pipeline A can expose TCP port 1600 and Pipeline B can expose UDP port 1600. Similarly, Pipeline C can’t have two different plugins that expose TCP port 1850, but it can have one plugin that exposes TCP port 1850 and another plugin that exposes TCP port 1851.

Automatic reassignment

If two pipeline components within a Core Instance try to claim the same port, Telemetry Pipeline resolves this conflict by reassigning the frontend port of the newer component to a different port number.

This reassignment process increments the frontend port number of the newer component by one until it finds a non-conflicting port. For example, if Pipeline B has a plugin whose frontend and backend port are set to TCP port 2200, but that port number is already in use by Pipeline A, Telemetry Pipeline will attempt to reassign the frontend port of Pipeline B’s plugin to TCP port 2201. If TCP port 2201 is also already in use, Telemetry Pipeline will attempt to reassign the frontend port to TCP port 2202 instead.

As a result of this reassignment process, a component can have a different frontend port number and backend port number. In the previous example, if Telemetry Pipeline successfully reassigned Pipeline B’s plugin on the first attempt, then that plugin’s frontend port number would be 2201 and its backend port number would still be 2200.

Proxies

You can use an HTTP proxy to route traffic between a Core Instance and any connected services. To do so, use the following command:

calyptia update core_instance operator INSTANCE --http-proxy=PROXY

Replace the following values:

• INSTANCE: The name of your Core Instance.
• PROXY: The proxy to use (for example, http://localhost:3128).

You can also specify an HTTPS proxy with the https-proxy flag.

Service types

Each pipeline uses one of the available Kubernetes Service (opens in a new tab) types. The Service types supported by Telemetry Pipeline are:

• ClusterIP (opens in a new tab)
• LoadBalancer (opens in a new tab)
• NodePort (opens in a new tab)

Create a pipeline with a specific Service type

When creating a pipeline, use the --service-type flag to specify its Service type:

calyptia create pipeline --name PIPELINE --core-instance INSTANCE --service-type TYPE

Replace the following values:

• PIPELINE: A short name you’d like to give your pipeline.
• INSTANCE: The name of your Core Instance.
• TYPE: The Service type for your pipeline (ClusterIP, LoadBalancer, or NodePort).

Modify the Service type of an existing pipeline

To modify the Service type of an existing pipeline, use the following command:

calyptia update pipeline PIPELINE --service-type TYPE

Replace the following values:

• PIPELINE: The name or unique ID of your existing pipeline.
• TYPE: The Service type for your pipeline (ClusterIP, LoadBalancer, or NodePort).