OpenTelemetry’s march towards beta

OpenTelemetry has a large presence at Kubecon this year. This post serves as a project update for the conference and describes the milestones ahead of us in the next year.

OpenTelemetry, a project that aims to make robust, portable telemetry a built-in feature of cloud-native software, will reach beta and GA status early next year.

OpenTelemetry’s interfaces and libraries for Go, Java, JavaScript, Python, .Net, C++, Erlang, PHP, and Ruby provide the components needed to capture distributed traces, application metrics, and other critical signals from their code. These include:

• APIs for defining and controlling the behaviour of traces, metrics, and metadata
• An SDK for each language that implements the APIs and manages context, views, sampling, buffering, etc.
• A growing list of integrations with web framework, RPC systems, storage clients, etc. that allow the libraries to automatically propagate context and capture traces and metrics with no developer interaction beyond instantiation
• Exporters that send captured signals to developers’ backends of choice, including Zipkin, Jaeger, Prometheus, and commercial APM options
• A collector that can capture system metrics and receive and re-export telemetry captured by OpenTelemetry libraries, as well as Prometheus, Jaeger, and Zipkin clients

OpenTelemetry is a key part of instrumenting your applications for observability:

• OpenTelemetry provides the components that developers and operators of cloud-native services need to capture distributed traces, application metrics, and other critical signals from their code. It allows data to be exported to any supported backends without changing your code.
• Maintainers of clients for web frameworks, databases, etc. who want to provide better observability and debuggability to their users can start integrating with each language’s API as it reaches beta. These integrations will allow your customers to capture the telemetry that you define by simply setting up OpenTelemetry for their applications.
• Vendors of APM (Application Performance Management) products, observability products, and cloud platforms are already working to support OpenTelemetry in their offerings.

After the core libraries and agent reach beta over the coming weeks, the project will collect feedback from early adopters, collect benchmark and other performance data on the SDK, and work towards beta and general availability releases in the first quarter of 2020.

Want to join us? Check out this recent post from OpenTelemetry contributor Austin Parker for more insight on how to get started contributing to the project today.