OpenTelemetry Support

Using OpenTelemetry with Sentry Performance.

You can configure your OpenTelemetry SDK to send traces and spans to Sentry.

The sentry-opentelemetry gem requires opentelemetry-sdk 1.0.0 or higher.

gem "sentry-ruby"
gem "sentry-rails"
gem "sentry-opentelemetry"

gem "opentelemetry-sdk"
gem "opentelemetry-instrumentation-all"

Initialize Sentry for tracing and set the instrumenter to :otel:

Sentry.init do |config|
  config.dsn = ""
  config.traces_sample_rate = 1.0
  # set the instrumenter to use OpenTelemetry instead of Sentry
  config.instrumenter = :otel

This disables all Sentry instrumentation and relies on the chosen OpenTelemetry tracers for creating spans.

Next, configure OpenTelemetry as you need and hook in the Sentry span processor and propagator:

OpenTelemetry::SDK.configure do |c|

OpenTelemetry.propagation =

Note that if you need extra Sentry features such as tags / user, you will need to re-order the middleware as below to make sure Sentry's scope management works properly.


With Sentry’s OpenTelemetry SDK, an OpenTelemetry Span becomes a Sentry Transaction or Span. The first Span sent through the Sentry SpanProcessor is a Transaction, and any child Span gets attached to the first Transaction upon checking the parent Span context. This is true for the OpenTelemetry root Span and any top level Span in the system. For example, a request sent from frontend to backend will create an OpenTelemetry root Span with a corresponding Sentry Transaction. The backend request will create a new Sentry Transaction for the OpenTelemetry Span. The Sentry Transaction and Span are linked as a trace for navigation and error tracking purposes.

If you need more fine grained control over Sentry, take a look at the Configuration page. In case you'd like to filter out transactions before sending them to Sentry (to get rid of health checks, for example), you may find the Filtering page helpful.

Help improve this content
Our documentation is open source and available on GitHub. Your contributions are welcome, whether fixing a typo (drat!) or suggesting an update ("yeah, this would be better").