O-RAN Telemetry Demo with Expanso on OpenShift

Demo showing Expanso Edge pipeline consuming O-RAN metrics and pushing to OTEL Collector on OpenShift.

Landing Page | Quick Start Guide

Overview

This demo showcases how Expanso can collect, transform, and route O-RAN telemetry data from edge Distributed Units (DUs) running on OpenShift Single Node deployments. The pipeline transforms raw DU metrics into OpenTelemetry Protocol (OTLP) format for ingestion by any OTLP-compatible observability backend.

Key Capabilities

Edge-native processing - Runs directly on SNO nodes with minimal footprint
Real-time transformation - Bloblang processors normalize and enrich telemetry
Multi-destination routing - Fan-out to multiple OTLP endpoints simultaneously
Cloud-managed pipelines - Deploy and update pipelines via Expanso Cloud

Red Hat Integration

This demo is designed for integration with Red Hat's observability stack:

OpenShift Container Platform (Single Node deployment)
OpenTelemetry Collector (metrics routing)
Red Hat Observability or any OTLP-compatible backend

Expanso Pipeline Architecture

Expanso Edge pipelines follow an Input → Processors → Output model:

┌─────────────────────────────────────────────────────────────────┐
│                      Expanso Edge Pipeline                      │
│                                                                 │
│   ┌─────────┐     ┌─────────────────────┐     ┌─────────────┐  │
│   │  INPUT  │────▶│     PROCESSORS      │────▶│   OUTPUT    │  │
│   │         │     │                     │     │             │  │
│   │ - file  │     │ - transform         │     │ - HTTP      │  │
│   │ - http  │     │ - filter            │     │ - OTLP      │  │
│   │ - kafka │     │ - enrich            │     │ - Kafka     │  │
│   │ - generate│   │ - validate          │     │ - stdout    │  │
│   └─────────┘     └─────────────────────┘     └─────────────┘  │
│                                                                 │
└─────────────────────────────────────────────────────────────────┘

Pipeline Configuration (Bloblang)

The pipeline uses Bloblang for data transformation. Key patterns:

pipeline:
  processors:
    - mapping: |
        # Transform raw metrics to OTLP format
        let ts = (this.timestamp * 1000000000).string()

        root.resourceMetrics = [{
          "resource": {
            "attributes": [
              {"key": "service.name", "value": {"stringValue": "du-simulator"}},
              {"key": "du.id", "value": {"stringValue": this.du_id}}
            ]
          },
          "scopeMetrics": [{
            "scope": {"name": "du-telemetry", "version": "1.0.0"},
            "metrics": [
              {
                "name": "du.ptp4l_offset_ns",
                "unit": "ns",
                "gauge": {
                  "dataPoints": [{"timeUnixNano": $ts, "asInt": this.ptp4l_offset_ns}]
                }
              }
            ]
          }]
        }]

Demo Architecture

┌──────────────────────────────────────────────────────────────────────────────┐
│                        OpenShift Single Node (Edge Site)                     │
│                                                                              │
│  ┌─────────────────┐     ┌─────────────────┐     ┌─────────────────┐        │
│  │  O-RAN Metrics  │────▶│  Expanso Edge   │────▶│  OTEL Collector │        │
│  │  (DU Simulator) │     │  (Pipeline)     │     │                 │        │
│  └─────────────────┘     └─────────────────┘     └────────┬────────┘        │
│         INPUT              PROCESSORS              OUTPUT  │                 │
│                            - transform to OTLP            │                  │
│                                                           ▼                  │
│                          ┌─────────────────┐     ┌─────────────────┐        │
│                          │     Grafana     │◀────│   Prometheus    │        │
│                          └─────────────────┘     └─────────────────┘        │
│                                                                              │
└──────────────────────────────────────────────────────────────────────────────┘

Data Flow

Input: DU simulator generates O-RAN telemetry every 5 seconds
Processing: Bloblang transforms metrics to OTLP JSON format
Output: HTTP POST to OTEL Collector at /v1/metrics
Storage: Prometheus scrapes OTEL Collector's Prometheus exporter
Visualization: Grafana queries Prometheus for dashboards

Multi-Destination Output

The pipeline can fan-out to multiple OTLP Collectors:

┌──────────────────────────────────────────────────────────────────────────────┐
│                        OpenShift Single Node (Edge Site)                     │
│                                                                              │
│  ┌─────────────────┐     ┌─────────────────┐     ┌─────────────────┐        │
│  │  O-RAN Metrics  │────▶│  Expanso Edge   │──┬─▶│  OTEL Collector │        │
│  │  (DU Simulator) │     │  (Pipeline)     │  │  │  (local)        │        │
│  └─────────────────┘     └─────────────────┘  │  └─────────────────┘        │
│         INPUT              PROCESSORS         │                              │
│                                               │                              │
└───────────────────────────────────────────────┼──────────────────────────────┘
                                                │
                                                │ OUTPUT (multiple destinations)
                                                ▼
                                      ┌─────────────────┐
                                      │  OTEL Collector │
                                      │  (external/cloud)│
                                      └─────────────────┘

This enables:

Local observability for on-site operators
Central aggregation for fleet-wide visibility
Hybrid routing based on metric criticality

O-RAN Metrics

The demo simulates key O-RAN DU telemetry:

Metric	Unit	Description
`du.ptp4l_offset_ns`	nanoseconds	PTP clock offset from grandmaster
`du.cpu_pct`	percent	CPU utilization on isolated cores
`du.prb_dl_pct`	percent	Physical Resource Block utilization (downlink)

PTP Compliance Thresholds

Per O-RAN and Red Hat RAN Reference Design specifications:

Metric	Threshold	Status
ptp4l offset	±100ns	DEGRADED if exceeded
phc2sys offset	±50ns	DEGRADED if exceeded
Lock state	!= LOCKED	CRITICAL

Prerequisites

OpenShift cluster - Single Node OpenShift (SNO) or standard cluster
oc CLI - Logged in with permissions to create deployments, services, routes
Expanso Cloud account - Register at cloud.expanso.io

OpenShift Requirements

# Verify cluster access
oc whoami
oc get nodes

# Ensure you have a project/namespace
oc project <your-project>

Deployment

Step 1: Register Node with Expanso Cloud

Create your Expanso Cloud account and set up credentials for the edge node.

Go to cloud.expanso.io
Create a new Network (logical grouping of edge nodes)
Generate a bootstrap token - this authenticates the edge node

Note: The bootstrap token is a one-time credential. Store it securely.

Step 2: Create Bootstrap Secret

Store the bootstrap token as a Kubernetes secret:

oc create secret generic expanso-bootstrap \
  --from-literal=token=<BOOTSTRAP_TOKEN>

Verify the secret:

oc get secret expanso-bootstrap -o yaml

Step 3: Deploy Observability Stack

Deploy the OTEL Collector, Prometheus, and Grafana:

oc apply -f observability-deployment.yaml

This creates:

OTEL Collector - Receives OTLP metrics, exports to Prometheus format
Prometheus - Scrapes OTEL Collector, stores time-series data
Grafana - Visualization and dashboards

Verify pods are running:

oc get pods -w

Expected output:

NAME                              READY   STATUS    RESTARTS   AGE
otel-collector-xxxxxxxxx-xxxxx    1/1     Running   0          30s
prometheus-xxxxxxxxx-xxxxx        1/1     Running   0          30s
grafana-xxxxxxxxx-xxxxx           1/1     Running   0          30s

Step 4: Deploy Expanso Edge

Deploy the Expanso Edge agent:

oc apply -f expanso-edge-deployment.yaml

The agent will:

Read the bootstrap token from the secret
Register itself with Expanso Cloud
Begin polling for pipeline configurations

Verify registration:

oc logs -f deployment/expanso-edge

Step 5: Apply Pipeline via Expanso Cloud

Deploy the pipeline through Expanso Cloud:

Go to cloud.expanso.io
Navigate to your Network
Click Create Pipeline
Paste the contents of expanso-pipeline-FOR-EXPANSO-CLOUD.yaml
Click Deploy

The pipeline will be pushed to all nodes in the network.

Step 6: View Metrics in Grafana

Access the Grafana dashboard:

# Get the route URL
oc get route grafana -o jsonpath='{.spec.host}'

Open the URL in your browser
Login with admin / admin
Go to Explore → Select Prometheus datasource
Query metrics:
- du_ptp4l_offset_ns
- du_cpu_pct
- du_prb_dl_pct

Files Reference

File	Description
`observability-deployment.yaml`	OTEL Collector, Prometheus, Grafana stack
`expanso-edge-deployment.yaml`	Expanso Edge agent deployment
`expanso-pipeline-FOR-EXPANSO-CLOUD.yaml`	Pipeline config (for Expanso Cloud UI)
`expanso-pipeline-FOR-EXPANSO-CLI.yaml`	Pipeline config (for CLI deployment)

YAML Structure

observability-deployment.yaml:

# ConfigMaps for OTEL Collector and Prometheus
# Deployments for all three services
# Services for internal communication
# Route for Grafana external access

expanso-edge-deployment.yaml:

# Deployment with bootstrap token from secret
# Service for Expanso Edge API (port 9010)
# Route for external access (optional)

Troubleshooting

Pipeline not receiving data

# Check Expanso Edge logs
oc logs deployment/expanso-edge

# Check OTEL Collector logs
oc logs deployment/otel-collector

# Verify services are running
oc get svc

Metrics not appearing in Grafana

# Check Prometheus targets
oc port-forward svc/prometheus 9090:9090
# Open http://localhost:9090/targets

# Verify OTEL Collector is exporting
oc port-forward svc/otel-collector 8889:8889
curl http://localhost:8889/metrics

Bootstrap token issues

# Verify secret exists
oc get secret expanso-bootstrap

# Check token value (base64 encoded)
oc get secret expanso-bootstrap -o jsonpath='{.data.token}' | base64 -d

Cleanup

Remove all deployed resources:

oc delete -f expanso-edge-deployment.yaml
oc delete -f observability-deployment.yaml
oc delete secret expanso-bootstrap

Name		Name	Last commit message	Last commit date
Latest commit History 21 Commits
.claude		.claude
.github/workflows		.github/workflows
docs		docs
.cursorindexingignore		.cursorindexingignore
.gitignore		.gitignore
.pre-commit-config.yaml		.pre-commit-config.yaml
.yamllint		.yamllint
README.md		README.md
expanso-edge-deployment.yaml		expanso-edge-deployment.yaml
expanso-pipeline-FOR-EXPANSO-CLI.yaml		expanso-pipeline-FOR-EXPANSO-CLI.yaml
expanso-pipeline-FOR-EXPANSO-CLOUD.yaml		expanso-pipeline-FOR-EXPANSO-CLOUD.yaml
observability-deployment.yaml		observability-deployment.yaml

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

O-RAN Telemetry Demo with Expanso on OpenShift

Overview

Key Capabilities

Red Hat Integration

Expanso Pipeline Architecture

Pipeline Configuration (Bloblang)

Demo Architecture

Data Flow

Multi-Destination Output

O-RAN Metrics

PTP Compliance Thresholds

Prerequisites

OpenShift Requirements

Deployment

Step 1: Register Node with Expanso Cloud

Step 2: Create Bootstrap Secret

Step 3: Deploy Observability Stack

Step 4: Deploy Expanso Edge

Step 5: Apply Pipeline via Expanso Cloud

Step 6: View Metrics in Grafana

Files Reference

YAML Structure

Troubleshooting

Pipeline not receiving data

Metrics not appearing in Grafana

Bootstrap token issues

Cleanup

Learn More

About

Uh oh!

Releases

Packages

Contributors 3

Uh oh!

aronchick/oran-sample

Folders and files

Latest commit

History

Repository files navigation

O-RAN Telemetry Demo with Expanso on OpenShift

Overview

Key Capabilities

Red Hat Integration

Expanso Pipeline Architecture

Pipeline Configuration (Bloblang)

Demo Architecture

Data Flow

Multi-Destination Output

O-RAN Metrics

PTP Compliance Thresholds

Prerequisites

OpenShift Requirements

Deployment

Step 1: Register Node with Expanso Cloud

Step 2: Create Bootstrap Secret

Step 3: Deploy Observability Stack

Step 4: Deploy Expanso Edge

Step 5: Apply Pipeline via Expanso Cloud

Step 6: View Metrics in Grafana

Files Reference

YAML Structure

Troubleshooting

Pipeline not receiving data

Metrics not appearing in Grafana

Bootstrap token issues

Cleanup

Learn More

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Contributors 3

Uh oh!

Packages