OpenTelemetry Export

GitLab Pipelines in Your Observability Stack

Export GitLab CI/CD pipelines as standards-compliant OTLP traces. Correlate deployments with production incidents, monitor pipeline performance, and get complete observability from request to deployment.

Standards-Compliant Zero Code Changes Works Everywhere
Honeycomb Trace 200 OK
acme/api-service
cicd.pipeline.run.id: 12345
cicd.pipeline.run.result: success
vcs.ref: main
4m 32s
build
1m 45s
test
2m 12s
deploy
35s

Pipeline Data Trapped in Silos

Your deployments happen in GitLab. Your monitoring lives in Datadog or Honeycomb. When production breaks, you're hunting across multiple tools trying to connect the dots.

No Deployment Correlation

Production error spike at 2:35 PM? Good luck figuring out which pipeline deployed the bad code. You're manually cross-referencing timestamps between GitLab and your APM.

Invisible to SRE Teams

Your SRE team lives in Grafana. They have no visibility into deployments without switching to GitLab. CI/CD is a black box in their monitoring dashboards.

No Historical Metrics

Want to know if your pipelines are getting slower? Build custom scripts to scrape GitLab's API, parse the data, and ship it somewhere. There's no built-in way to get CI/CD metrics into your observability stack.

Manual Incident Investigation

Distributed tracing shows the request flow through your services. But where did that service version come from? You're clicking through GitLab trying to find the deployment that matches the timestamp.

Standards-Compliant OTLP Trace Export

Every GitLab pipeline becomes an OpenTelemetry trace. Automatically exported to your observability backend with full CI/CD semantic conventions. No code changes, no custom scripts, just standard OTLP.

Automatic Traces

Every completed pipeline automatically becomes an OTLP trace with pipeline and job spans. Parent-child relationships preserved.

Semantic Conventions

Full compliance with OpenTelemetry CI/CD semantic conventions. Standard attributes like cicd.pipeline.name and cicd.pipeline.task.type.

Dual Protocols

HTTP/JSON and HTTP/Protobuf support. Works with Honeycomb, Datadog, Grafana Tempo, and any OTLP-compatible collector.

Deterministic IDs

Trace IDs derived from GitLab pipeline IDs. Same pipeline = same trace ID. Perfect for correlation across systems.

How It Works

Pipie sits between GitLab and your observability backend, transforming webhook events into standards-compliant OTLP traces.

GitLab CI/CD

Pipeline and job webhooks

Pipie OTLP Export

Converts to OTLP with semantic conventions

Your OTEL Backend

Honeycomb, Datadog, Grafana, etc.

Example Trace Attributes

Pipeline Span
cicd.pipeline.name = "acme/api-service"
cicd.pipeline.run.id = "12345"
cicd.pipeline.run.result = "success"
vcs.ref = "main"
vcs.revision = "abc123def456"
Job/Task Span
cicd.pipeline.task.name = "test"
cicd.pipeline.task.type = "test"
cicd.pipeline.task.run.id = "67890"
cicd.pipeline.task.run.result = "success"

Works With Your Existing Stack

Standards-compliant OTLP means it works with any OpenTelemetry-compatible backend. Popular platforms, self-hosted collectors, or your custom solution.

Honeycomb

Native OTLP support with powerful query language for CI/CD trace analysis.

  • Native OTLP endpoint
  • Trace waterfall views
  • SLO monitoring for pipelines

Datadog

Unified observability platform with CI Visibility and APM correlation.

  • CI Visibility integration
  • APM trace correlation
  • Custom dashboards & alerts

Grafana / Tempo

Open-source stack with Tempo for trace storage and Grafana for visualization.

  • Self-hosted option
  • Grafana dashboards
  • Prometheus metrics correlation
AND ANY OTLP-COMPATIBLE BACKEND:
New Relic Dynatrace Elastic Splunk

Real-World Use Cases

From incident investigation to DORA metrics, OTLP trace export unlocks powerful workflows.

Deployment Correlation

Trace production errors back to the exact pipeline that deployed the code. Click from your OTEL backend directly to the GitLab pipeline.

Reduce incident investigation time by 80%

Performance Monitoring

Track pipeline duration trends over time. Identify slow stages, measure improvement, and get alerted when builds slow down.

Monitor CI/CD performance without custom tooling

Incident Investigation

Follow distributed traces from production requests back through the CI/CD pipeline that built and deployed the service.

Complete observability from request to deployment

Team Visibility

SRE teams see deployments in their existing dashboards. No context switching between GitLab and observability tools.

Unified visibility across your entire stack

DORA Metrics

Calculate deployment frequency, lead time, and failure rates from OTEL traces. Export to your business intelligence tools.

Engineering metrics from standard telemetry

5-Minute Setup

No code changes. No custom scripts. Just configure and start exporting traces.

  1. Enable OTEL Export
    Toggle OTEL export in settings - no code changes required
    ⚡ Enable
  2. Configure Endpoint
    Paste your Honeycomb, Datadog, or OTEL collector endpoint URL
    🔗 Endpoint
  3. Add Authentication
    Configure API key or custom headers for your observability backend
    🔑 Auth
  4. Test Connection
    Send a test trace to verify everything is working
    ✅ Test
  5. Go Live
    Pipeline traces automatically exported on completion
    🚀 Live

Technical Specifications

Everything you need to know about our OpenTelemetry implementation.

What OTLP version is supported?
We support OTLP v1.0.0 with full CI/CD semantic conventions as defined by the OpenTelemetry specification.
Which protocols are available?
Both HTTP/JSON (application/json) and HTTP/Protobuf (application/x-protobuf) protocols are fully supported. Choose based on your backend's requirements.
When are traces exported?
Traces are exported only when pipelines and jobs reach terminal states (success, failed, canceled, skipped). This prevents duplicate traces and ensures accurate duration measurements.
How are trace IDs generated?
Trace IDs are deterministically generated from GitLab pipeline IDs, ensuring the same pipeline always produces the same trace ID. This enables perfect correlation across systems.
What about security?
Custom headers (including API keys) are encrypted at rest. All exports use HTTPS. We never store your trace data - it goes directly to your configured endpoint.
How does retry logic work?
Transient failures (429, 502, 503, 504) are automatically retried with exponential backoff, up to 3 attempts. Permanent failures (400, 401) are not retried.

OpenTelemetry CI/CD Semantic Conventions

We follow the official OpenTelemetry CI/CD semantic conventions for all attributes:

Pipeline Attributes

  • • cicd.pipeline.name
  • • cicd.pipeline.run.id
  • • cicd.pipeline.run.result
  • • cicd.pipeline.run.state
  • • cicd.pipeline.run.url.full

Task/Job Attributes

  • • cicd.pipeline.task.name
  • • cicd.pipeline.task.type
  • • cicd.pipeline.task.run.id
  • • cicd.pipeline.task.run.result
  • • cicd.pipeline.task.run.url.full

Included in All Paid Plans

OpenTelemetry export is available on all paid Pipie plans. Start with a 14-day free trial.

Pricing

Plans that scale with your team

Choose a plan that fits your team size and workflow needs. Start free and upgrade as you grow.

Pro Plan

Perfect for growing teams that need smart notifications and reviewer matching.

$30 /month

Start Trial
  • Up to 15 contributors
  • Merge request tracking
  • Pipeline notifications
  • $1.50 per extra contributor

Business Plus Plan

Advanced features for large teams with complex workflows and compliance needs.

$150 /month

Start Trial
  • Up to 100 contributors
  • Priority support
  • Merge request tracking
  • Pipeline notifications
  • Team collaboration tools
  • $1 per extra contributor

Self-Hosted

Deploy on your own infrastructure with complete data control.

Custom

Contact Us
  • Deploy on your servers
  • Complete data sovereignty
  • Custom SLAs available
  • Unlimited contributors

Start Exporting Traces in 5 Minutes

Join engineering teams using Pipie to get complete observability from GitLab to production. No credit card required for trial.

14-day free trial
No credit card required
Cancel anytime