OpenTelemetry and Observability: Looking Forward
As we approach the end of the year, it’s a wonderful time to pause and reflect. 2023 was a landmark for OpenTelemetry, as its three fundamental signals, tracing, metrics and logging reached stable releases. This achievement signifies the realization of OpenTelemetry’s original vision to provide a standards-based framework for instrumenting and collecting observability data.
Let’s take this opportunity to explore some of the exciting trends we’ve witnessed, delve into innovative products and use cases, and thoughtfully consider the evolving landscape of observability as we anticipate what 2024 has in store.
Metrics Came into its Own
While the OpenTelemetry specifications for metrics were declared stable in May 2022, this year has seen the adoption spread. Here are a couple of articles from practitioners:
- “Experience Report: Adopting OpenTelemetry for Metrics in Cloud Foundry” by Matthew Kocher and Carson Long from VMware.
- “Adding OpenTelemetry Metrics in Your Go App” by our very own Matheus Nogueira.
Looking forward to 2024, expect to see the same type of movement and adoption of logs.
Focus on Using Distributed Tracing in Load Tests
Two of the leading load testing tools, Grafana k6 and Artillery.io, added support for OpenTelemetry in 2023.
- Grafana k6 introduced tracing capabilities, enabling performance engineers to identify system bottlenecks or failures during load tests.
- Artillery.io followed, adding both metrics and distributed tracing, offering a more detailed analysis of system performance.
Tracetest leverages the capability exposed by k6 tests to enable trace-based tests of load tests, allowing deep assertions as you run the tests. We have seen widespread use of this capability by customers such as Sigma Software. In 2024, the Tracetest team will look at adding this same capability to Artillery.io and other load-testing tools.
Support and Use Cases for OpenTelemetry Expands
More vendors are embracing OpenTelemetry standards to support actions outside of the typical, but very vital, role of analyzing your telemetry data.
- Companies such as Tyk are instrumenting their API gateway so it natively supports OpenTelemetry.
- End users are finding new use cases for OpenTelemetry, such as using distributed tracing to observe your CI/CD Pipelines.
- Tracetest leverages distributed tracing data for integration and end-to-end tests.
Emphasis on the OpenTelemetry Collector
The OpenTelemetry Collector sits in the middle of the OpenTelemetry world, receiving signals from the application, processing and transforming them, and then exporting them to any number of back-end systems. As integrations and vendor support for OpenTelemetry expands, the needs and demands for this centralized collector increase.
The introduction of OpenTelemetry Transformation Language (OTTL) in 2023 added to the OpenTelemetry Collector’s ability to process and transform incoming signals.
At Tracetest, we were able to leverage the ability to use OTTL in the filter processor to improve the way we can gather trace data from production environments outputting large volumes of telemetry data. This change to the OpenTelemetry Collector’s filter processor makes Tracetest suitable for running tests in high-volume environments, including production.
Observability Everywhere
In recent discussions, we encountered a growing trend among customers toward an “observability everywhere” approach. Moving beyond its traditional use by site reliability engineers and DevOps, these companies involve everyone, including developers and testers, in observability. This shift redefines observability from a reactive tool for production issues to a proactive tool beneficial in development and testing.
Honeycomb has emphasized using observability during development, and tools such as Digma.ai and Tracetest are leading the way in this push forward.
The Browser
OpenTelemetry’s main role has been relegated to instrumenting back-end systems, with open, standards-based browser instrumentation still experimental and lagging. Work is being done on improving and standardizing this instrumentation.
- Uzufly stands out in this context. It uses existing client instrumentation to build tests. Looking ahead, its ambition is to expand trace-based testing to cover tests initiated by frontend actions within a browser.
This would enable full end-to-end tests across both the front end and back end. Stay tuned in 2024 for more on this topic!
2023 is ‘So Last Year’
Bidding goodbye to 2023, we look forward to 2024 with enthusiasm. OpenTelemetry has momentum, backed by standards and widespread adoption, powering its growth. The new year promises exciting developments, with innovative products and use cases emerging around OpenTelemetry. I am eager to see the advancements and innovations that 2024 will unveil. Long live OpenTelemetry!
