Jaeger: The Future with OpenTelemetry and Metrics

The presentation discusses the introduction of new signals, metrics, in addition to traces in Jaeger, a distributed tracing system, to understand performance and service monitoring. The auto-instrumentation feature in .NET eliminates the need for code changes to instrument the application.

• Jaeger is a distributed tracing system that helps debug and understand transactions
• Metrics are introduced as new signals in addition to traces to understand performance and service monitoring
• The span map metrics processor is used to derive metrics from traces and generate metrics from these traces
• Metrics can be visualized in Grafana and used in the Jaeger UI
• Auto-instrumentation feature in .NET eliminates the need for code changes to instrument the application

The speaker mentions that the auto-instrumentation feature in .NET eliminates the need for code changes to instrument the application. This feature was introduced to help companies that use .NET APIs and were trying to set up Jaeger but did not have the auto-instrumentation feature at the time. With this feature, developers can set the environment variables via config maps to instrument the app without having to pull in some type of package.

In this session, we will start with an introduction to the Jaeger distributed tracing project and the basics of distributed tracing. Jaeger recently deprecated its native clients in favor of the clients maintained by the upcoming OpenTelemetry project. We will explain what this means to you as users and why we are changing the path forward. To help facilitate this transition, we will cover auto instrumentation of an application using an OpenTelemetry client and we will cover multiple best practices to build a scalable trace pipeline to deliver this data to a Jaeger backend. Moving Jaeger from a tracing system to a monitoring system has been the big push for the project in the last year. Made possible by OpenTelemetry and the processor layer which allows for the creation of metrics derived from traces in the pipeline. Operational monitoring is now possible using the new monitoring tab, which adds metrics capabilities to Jaeger UI via another graduated project, Prometheus. We are always seeking new collaborators, contributors, and users. We need your help, please join us!