Important: This documentation is about an older version. It's relevant only to the release noted, many of the features and functions have been updated or replaced. Please view the current version.
Span profiles with Traces to profiles for Java
Span Profiles represents a major shift in profiling methodology, enabling deeper analysis of both tracing and profiling data. Traditional continuous profiling provides an application-wide view over fixed intervals. In contrast, Span Profiles delivers focused, dynamic analysis on specific execution scopes within applications, such as individual requests or specific trace spans.
This shift enables a more granular view of performance, enhancing the utility of profiles by linking them directly with traces for a comprehensive understanding of application behavior. As a result, engineering teams can more efficiently identify and address performance bottlenecks.
To learn more about Span Profiles, refer to Combining tracing and profiling for enhanced observability: Introducing Span Profiles.
Pyroscope integrates with distributed tracing systems supporting the OpenTelemetry standard. This integration lets you link traces with the profiling data and find resource usage for specific lines of code for your trace spans.
Note
- Only CPU profiling is supported at the moment.
- Because of how sampling profilers work, spans shorter than the sample interval may not be captured.
To use Span Profiles, you need to:
- Configure Pyroscope to send profiling data
- Configure a client-side package to link traces and profiles: Java
- Configure the Tempo data source in Grafana or Grafana Cloud to discover linked traces and profiles
Before you begin
Your applications must be instrumented for profiling and tracing before you can use span profiles.
- Profiling: Your application must be instrumented with Pyroscope’s Java client SDK. Refer to the Java guide for instructions.
- Tracing: Your application must be instrumented with OpenTelemetry traces. Refer to the OpenTelemetry guide for instructions.
Configure the otel-profiling-java package
To start collecting Span Profiles for your Java application, you need to include otel-profiling-java as an extension to your application.
Assuming you have this sample application Docker image:
# [...]
EXPOSE 5000
CMD ["java", "-Dserver.port=5000", "-jar", "./my-app.jar" ]
By adding the OTel Java agent and the Pyroscope OTel Java Agent extension, you can enrich your profiles with span IDs. This makes it possible to query for span-specific profiling data in Grafana Tempo:
# [...]
EXPOSE 5000
## Add required libararies
ADD https://github.com/open-telemetry/opentelemetry-java-instrumentation/releases/download/v1.17.0/opentelemetry-javaagent.jar opentelemetry-javaagent.jar
ADD https://repo1.maven.org/maven2/io/pyroscope/otel/0.10.1.11/otel-0.10.1.11.jar pyroscope-otel.jar
ENV PYROSCOPE_APPLICATION_NAME=my-app
ENV PYROSCOPE_FORMAT=jfr
ENV PYROSCOPE_PROFILING_INTERVAL=10ms
ENV PYROSCOPE_PROFILER_EVENT=itimer
ENV PYROSCOPE_PROFILER_LOCK=10ms
ENV PYROSCOPE_PROFILER_ALLOC=512k
ENV PYROSCOPE_UPLOAD_INTERVAL=15s
ENV OTEL_JAVAAGENT_EXTENSIONS=./pyroscope-otel.jar
ENV OTEL_PYROSCOPE_ADD_PROFILE_URL=false
ENV OTEL_PYROSCOPE_ADD_PROFILE_BASELINE_URL=false
ENV OTEL_PYROSCOPE_START_PROFILING=true
## Useful for debugging
# ENV PYROSCOPE_LOG_LEVEL=debug
## Those environment variables need to be overwritten at runtime, if you are using Grafana Cloud
ENV PYROSCOPE_SERVER_ADDRESS=http://localhost:4040
# ENV PYROSCOPE_BASIC_AUTH_USER=123 ## Grafana Cloud Username
# ENV PYROSCOPE_BASIC_AUTH_PASSWORD=glc_secret ## Grafana Cloud Password / API Token
## Add the pyroscope and the opentelemetry java-agents
CMD ["java", "-Dserver.port=5000", "-javaagent:./opentelemetry-javaagent.jar", "-jar", "./my-app.jar" ]
Available configuration options
Flag | Description | Default |
---|---|---|
otel.pyroscope.start.profiling | Boolean flag to start PyroscopeAgent. Set to false if you want to start the PyroscopeAgent manually. | true |
otel.pyroscope.root.span.only | Boolean flag. When enabled, the tracer will annotate only the first span created locally (the root span), but the profile will include samples of all the nested spans. This may be helpful in case if the trace consists of multiple spans shorter than 10ms and profiler can’t collect and annotate samples properly. | true |
otel.pyroscope.add.span.name | Boolean flag. Controls whether the span name added to profile labels. | true |
View the span profiles in Grafana Tempo
To view the span profiles in Grafana Tempo, you need to have a Grafana instance running and a data source configured to link trace spans and profiles.
Refer to the data source configuration documentation to see how to configure the visualization to link trace spans with profiles.
To use a simple configuration, follow these steps:
Select a Pyroscope data source from the Data source drop-down.
Optional: Choose any tags to use in the query. If left blank, the default values of
service.name
andservice.namespace
are used.
The tags you configure must be present in the spans attributes or resources for a trace to profiles span link to appear. You can optionally configure a new name for the tag. This is useful for example if the tag has dots in the name and the target data source doesn’t allow using dots in labels. In that case you can for example remap service.name to service_name.
- Select one or more profile types to use in the query. Select the drop-down and choose options from the menu.
The profile type or app must be selected for the query to be valid. Grafana doesn’t show any data if the profile type or app isn’t selected when a query runs.
Examples
Check out the examples directory for a complete demo application that shows tracing integration features.