Monitor polyglot applications with OBI
Overview
OpenTelemetry eBPF Instrumentation (OBI) supports various programming languages, including Go, Java, Python, Node.js, and .NET. It intercepts application communications at the kernel and library function levels, parses protocol semantics, and emits standard OpenTelemetry telemetry data. OBI automatically collects observability data without any code modification, regardless of the HTTP framework, database, or large model provider.
For AI observability, OBI provides built-in protocol-level tracing for four major GenAI providers: OpenAI, Anthropic, Google Gemini, and Qwen. It automatically identifies LLM calls, extracts tool call information from responses, and traces rerank and vector retrieval operations, covering the core stages of the RAG pipeline.
OBI supports Linux amd64 and arm64 architectures and requires a kernel version of 5.8 or higher (4.18 or higher for RHEL-based systems). You can deploy OBI as a standalone process, in a Docker container, or as a Kubernetes DaemonSet.
Release notes
Version | Date | Changes | Impact |
0.10.3 | August 7, 2026 |
| No service impact. |
0.10.2 | July 28, 2026 |
| No service impact. |
0.10.1 | July 20, 2026 |
| No service impact. |
0.10.0 | July 3, 2026 |
| No service impact. |
0.9.5 | June 24, 2026 |
| No service impact. |
0.9.4 | June 10, 2026 | Adds support for integrating applications without restarting the OBI component. | No service impact. |
0.9.3 | June 3, 2026 | Adds a profiling collection component. | No service impact. |
0.9.2 | May 27, 2026 | Adds support for integrating applications through Cloud Monitor 2.0 Integration Center. | No service impact. |
0.9.1 | May 22, 2026 | Initial release. | N/A |