Recently, the OpenTelemetry Protocol (OTLP) 1.0.0 was released. OTLP specification describes telemetry data's encoding, transport, and delivery mechanism between telemetry sources, intermediate nodes such as collectors, and telemetry backends. It is a general-purpose telemetry data delivery protocol designed in the scope of the OpenTelemetry project.
OpenTelemetry (OTEL) is an open-source Cloud Native Computing Foundation (CNCF) project formed from the merger of the OpenCensus and OpenTracing projects. It is a vendor-neutral open-source Observability framework for instrumenting, generating, collecting, and exporting telemetry data such as traces, metrics, and logs. The framework provides a set of APIs, libraries, agents, and collector services for capturing distributed traces and metrics. In addition, its specification reached 1.0.0 version status earlier in 2021 and was reported by InfoQ.
OpenTelemetry facilitates the instrumentation of application code using APIs to generate telemetry data and seamlessly guide the collection of metrics across different SDK implementations. It offers language-specific integrations and libraries, such as the OTel SDK for Java, Golang, .NET, and Python, to enable developers to instrument their code and capture telemetry data. The telemetry data from these libraries is gathered and transmitted to the OpenTelemetry Collector, utilizing the OTLP for data exchange between the client and server. OTLP defines a serialization schema that closely adheres to the data models of traces, metrics, and logs.
Acting as a central repository, the OpenTelemetry Collector receives, processes, and exports the telemetry data collected from various sources, serving both as a local agent for individual applications and as a gateway for multiple applications.
Open Telemetry Diagram (Source: OpenTelemetry Documentation)
OTLP plays a role in how data is serialized, deserialized, and transported between networked services. This protocol aims to specify a serialization schema that closely adheres to the data models and solves problems with other telemetry protocols.
Martin Thwaites, a Developer Advocate at Honeycomb, told InfoQ:
OTLP is the beating heart of OpenTelemetry and makes it more powerful than anything before; moving to v1 is taking this to the next level. Already, most vendors have adapted to accepting data using the OTLP protocol; with V1, this gives some much-needed confidence and will hopefully allow the last few holdouts to add OTLP support.
In addition, he stated that:
This is important as it opens up more interoperability and less need for adding proprietary protocol libraries to our stacks, which can only ever be a win for people wanting to get more visibility into what's happening with their applications.
Lastly, the OpenTelemetry project is currently a CNCF incubation project.