5 min read
Doron Grinstein

How OpenTelemetry Works

Instrumenting your code with OpenTelemetry is essential for gaining valuable insights into your application’s behavior and ensuring efficient monitoring and troubleshooting.

As software systems become increasingly complex and distributed, the need for observability and understanding the performance and behavior of these systems has never been more critical. Fortunately, OpenTelemetry is available to assist us in overcoming numerous observability obstacles we encounter today. This open-source project tackles the growing need for observability by offering a standardized way to instrument cloud-native applications. Let’s explore OpenTelemetry, why you should consider using it, and how it can improve your system’s observability.

What is OpenTelemetry?

OpenTelemetry is a Cloud Native Computing Foundation (CNCF) project that standardizes the collection and management of observability data (logs, metrics, and traces) across multiple languages and tools. It provides a unified way to instrument code, allowing developers to collect telemetry data from the events that occur in their systems, helping them understand their software’s performance and behavior. It has become widely adopted, with many popular tools such as Datadog, Lightstep, Splunk, Jaeger, Loki and many others that natively support OpenTelemetry. With OpenTelemetry’s broad adaption, you can trust that your favorite observability tool will handle the data from your instrumented code.

Why is instrumentation important?

To effectively monitor and troubleshoot applications, three types of data are essential: metrics, logs, and traces. 

  • Metrics – provide information about the overall performance and health of your applications and systems. Metrics allow you to observe usage trends over time and see the impact of changes. This data offers a quantitative view of system performance.
  • Logs – are text-based records of events occurring within an application or system, they give you detailed information about specific events. Logs are helpful in identifying issues and track historical patterns. 
  • Traces – help you gain insights into the execution paths of requests and troubleshoot transactions in distributed systems. Traces provide valuable data on performance that facilitate detection of bottlenecks and performance-related problems. 

By instrumenting your code, you can collect this essential data, allowing you to identify and resolve issues efficiently. Choosing the right instrumentation library is crucial, as changing libraries later can be time-consuming and expensive. This is where OpenTelemetry proves useful. It provides a standard for instrumentation that reduces lock-in risks that are associated with a specific tool.

How to instrument your code with OpenTelemetry?

You can use either manual or automatic methods to instrument your code with OpenTelemetry.

Manual instrumentation involves explicitly adding tracing and metric collection code to your application’s source code. OpenTelemetry provides libraries in various programming languages, like Java, Go, Python, .Net and more, to make manual instrumentation straight forward. As an example,  the OpenTelemetry Python library currently offers support for more than 40 integrations, such as Django, Flask, Kafka, Redis, and MySQL, among others. 

Automatic instrumentation, on the other hand, involves using third-party libraries or frameworks to automatically generate tracing and metrics data without the need to modify your application’s source code. That is a fast and easy way to collect data from your applications, especially if you are using a well-known framework for which automatic instrumentation is supported. 

To start using OpenTelemetry, you can visit the OpenTelemetry  registry, find the appropriate library for your programming language, and follow the instructions for instrumenting your code. As the OpenTelemetry project is rapidly evolving, some languages may have more mature support for certain types of data (e.g., traces, metrics, or logs). You can check the status of support for your language at the https://opentelemetry.io/ website.

OpenTelemetry Collector (otel-collector)

An important component of the OpenTelemetry framework is the OpenTelemetry Collector, also known as otel-collector. It is a Golang-based process. It enables receiving, processing, and exporting telemetry data from various sources. The Collector acts as a vendor-agnostic proxy that can receive data in various formats, such as Jaeger, Prometheus, OTLP, and many others, and send it to one or more backends, including commercial/proprietary tools. With otel-collector you have the flexibility to change backends without altering instrumentation, creating a single set of standards for working with multiple vendors, projects, and platforms, and simplifying telemetry data management and export.

By using otel-collector you will reduce costs and improve performance because the collector can reduce the amount of data sent to the backend by filtering and transforming the data before sending it. Furthermore, when multiple applications transmit telemetry data in various formats, the otel-collector provides a common representation of the data to the downstream pipelines by translating between different data formats.


Instrumenting your code with OpenTelemetry is essential for gaining valuable insights into your application’s behavior and ensuring efficient monitoring and troubleshooting. Using OpenTelemetry reduces the risk of being locked into a particular tool and provides the flexibility to switch to another observability tool if needed. 

The Control Plane platform provides out of the box OpenTelemetry instrumentation for your apps, so that logs, metrics, tracing, and their aggregations across regions is automatic and painless. To learn more about Control Plane, please visit https://controlplane.com