Collecting GitHub Actions Traces with OpenTelemetry and SigNoz

GitHub Actions and similar CI/CD solutions are crucial for automation, maintainability, and project lifecycle management. They help enforce standards and keep things clean — but when it comes to workflow observability, what are we really talking about?

The Problem

Normally, when a DevOps or SRE engineer looks at GitHub Actions, they see:

• Individual actions and steps
• Workflows separated by repository
• Failures (maybe) — if someone reports them
• Performance issues — only when developers complain

But imagine managing a complex project with many GitHub workflows, hundreds of steps, and countless moving parts. How do you know:

• What are the bottlenecks?
• Where are the issues?
• What's causing latency?
• What's the success rate of a given action?

You need performance KPIs, metrics, and traces that tell you the health of your workflows. That's exactly what we're going to implement.

In this guide, I'm going to show you how to collect traces from your GitHub repository workflows and actions, and provide a dashboard with traces and metrics that will help you improve, detect anomalies early, and optimize your GitHub Actions workflows.

---

Architecture Overview

For this demo, I have the following environment setup:

• k3s cluster — lightweight Kubernetes
• SigNoz — installed in k3s
• OpenTelemetry Operator — installed and ready

Here's how the architecture works:

1. GitHub repository with a configured webhook pointing to a public URL (I'm using an ngrok URL for demo purposes)
2. ngrok exposing an OpenTelemetry Collector with GitHub receiver via NodePort service
3. OpenTelemetry Collector created by the OTel Collector CR, pointing to SigNoz as the exporter backend
4. SigNoz UI to visualize and analyze traces

---

Prerequisites

Before we start, make sure you have:

• A Kubernetes cluster (I'm using k3s)
• SigNoz installed and running
• OpenTelemetry Operator installed
• A GitHub repository with admin access
• ngrok or another tunneling solution for exposing the webhook endpoint

---

Step 1: Create a GitHub Personal Access Token

First, we need to create a Fine-grained Personal Access Token in GitHub. Even though we're not scraping in this setup, GitHub forces us to configure a scraper — this is mentioned in the OpenTelemetry Collector Contrib source code for the GitHub receiver.

Let's create the token:

1. Go to GitHub Settings → Developer settings → Personal access tokens → Fine-grained tokens
2. Click "Generate new token"
3. Give it a name like "GitHub Actions Observability"
4. Set expiration (I'm using 90 days for this demo)
5. Select the repository you want to monitor
6. Grant the following permissions:
   • actions: read — to read workflow information
   • metadata: read — to read repository metadata

Loading image...

Loading image...

---

Step 2: Setup OpenTelemetry Collector with GitHub Receiver

Now I'm going to create the OpenTelemetry Collector configuration with the GitHub receiver. This collector will receive webhook events from GitHub and convert them into traces.

Create the Secret

First, let's create a Kubernetes secret with the GitHub token and webhook secret:

Create the OpenTelemetry Collector

Now let's create the OpenTelemetry Collector Custom Resource:

Expose the Webhook Service

Now we need to expose the webhook service so GitHub can reach it. For this demo, I'm using a NodePort service:

Apply these resources:

---

Step 3: Expose the Webhook Endpoint

Since we're running this in a local k3s cluster, we need to expose the webhook service publicly so GitHub can reach it. I'm using ngrok for this demo.

---

Step 4: Configure GitHub Webhook

Now I'm going to configure the webhook in GitHub to send events to our OpenTelemetry Collector.

Create the Webhook

1. Go to your GitHub repository
2. Navigate to Settings → Webhooks
3. Click Add webhook
4. Configure the webhook:
   • Payload URL: https://your-ngrok-url.ngrok.io/events
   • Content type: application/json
   • Secret: The same secret you used in the Kubernetes secret (in my case, topsecret)
   • Which events: Select Let me select individual events
   • Enable:
     • workflow_run — to track workflow runs
     • workflow_job — to track individual job executions

Loading image...

Loading image...

Test the Webhook

After creating the webhook, GitHub will send a test delivery. You can check if it was received successfully:

Loading image...

---

Step 5: Test the Setup

Now let's test the entire setup. I'm going to trigger a workflow by creating a PR and merging it to main. Since we're in a knowledge base repo, I'll use a simple workflow action that updates the website via GitHub Pages.

When the PR is merged and the workflow starts in the main branch, we should see traces appearing in SigNoz.

Viewing Traces in SigNoz

After the workflow runs, navigate to SigNoz and check the traces:

Loading image...

Loading image...

Perfect! We've successfully set up observability for GitHub Actions workflows.

---

What We've Achieved

Now we can:

• Detect anomalies in workflow execution
• Identify bottlenecks in CI/CD pipelines
• Track performance metrics for each workflow
• Monitor success rates across different actions
• Get early warnings when workflows start to degrade

This observability setup will help you:

• Optimize slow workflows
• Debug failures faster
• Understand workflow dependencies
• Track improvements over time

---

Next Steps

In the next document, I'm going to show you how to enhance this setup to also collect GitHub workflow metrics. This is a simple addition to the collector configuration that will give you even more insights into your CI/CD pipeline performance.

---

Troubleshooting

Webhook Not Receiving Events

If you're not seeing traces in SigNoz:

1. Check the webhook delivery logs in GitHub
2. Verify the ngrok tunnel is still active
3. Check the OpenTelemetry Collector logs: kubectl logs -n opentelemetry-operator-system -l app.kubernetes.io/component=opentelemetry-collector
4. Ensure the webhook secret matches in both GitHub and the Kubernetes secret

Collector Not Starting

If the collector pod isn't running:

1. Check pod status: kubectl get pods -n opentelemetry-operator-system
2. Check pod logs: kubectl logs -n opentelemetry-operator-system <pod-name>
3. Verify the secret exists and has the correct keys
4. Check the OpenTelemetry Collector CR for syntax errors

No Traces in SigNoz

If traces aren't appearing in SigNoz:

1. Verify SigNoz is running: kubectl get pods -n signoz
2. Check the OTLP exporter endpoint is correct
3. Verify network connectivity between collector and SigNoz
4. Check the collector debug logs for export errors

---

Summary

We've successfully set up observability for GitHub Actions workflows using:

• OpenTelemetry Collector with GitHub receiver
• GitHub webhooks for real-time event delivery
• SigNoz for trace visualization and analysis

This setup gives you complete visibility into your CI/CD pipeline, helping you detect issues early and optimize performance. The traces show you exactly what's happening in your workflows, when it's happening, and how long each step takes.