What is the difference between Open Telemetry semantics and OpenInference semantics for AI observability?
Answered at 5:27 in Ep. #44, Beyond the Model: Observing AI Agents with Mikyo King · Heavybit Podcasts
What is the difference between Open Telemetry semantics and OpenInference semantics for AI observability?
OpenTelemetry is a general idea that applications have three primary signals: tracing (hierarchical interactions composed of spans that chain across network boundaries), structured logging, and metrics through Prometheus. With agents, you need to understand how they behave from the outside because code no longer just documents agent behavior—you don't know how many times it will loop or if it goes into code mode. OpenTelemetry specifically tracing can give understanding of that system from the outside, while semantics codify what that means (for example, db.xyz means something specific). Arize has been building those semantics over three years for agent-specific conventions.
Full answer and detailsRead the whole answer, key facts and every other question from this episode
In short
OpenTelemetry is a general idea that applications have three primary signals: tracing (hierarchical interactions composed of spans that chain across network boundaries), structured logging, and metrics through Prometheus.
Mikyo King, Head of Open Source at Arize · Ep. #44, Beyond the Model: Observing AI Agents with Mikyo King · Heavybit Podcasts · answer at 5:27
Key takeaways
- OpenTelemetry is a general idea that applications have three primary signals: tracing (hierarchical interactions composed of spans that chain across network boundaries), structured logging, and metrics through Prometheus.
Quote to remember
“It's one thing to capture the information. It's another thing to kind of make that the fossil fuel for improving your system as the next step.”
— Mikyo King
The full answer · about 4 min
OpenTelemetry is a general idea that applications have three primary signals: tracing (hierarchical interactions composed of spans that chain across network boundaries), structured logging, and metrics through Prometheus. With agents, you need to understand how they behave from the outside because code no longer just documents agent behavior—you don't know how many times it will loop or if it goes into code mode. OpenTelemetry specifically tracing can give understanding of that system from the outside, while semantics codify what that means (for example, db.xyz means something specific). Arize has been building those semantics over three years for agent-specific conventions.
Key facts
- Mikyo King has been a software developer for more than 20 years
- Last three years specifically focused on open source at Arize
- Arize started open source initiative around three years ago for observability and evaluation tooling
- Three years ago questions centered on LLM hallucinations, now focus shifted to agents as a service with higher levels of abstraction
People, companies and topics mentioned
More answers from this episode
- 3:26How does Arize use OpenTelemetry for AI agent observability?
- 13:44How should teams approach starting with evaluations for AI agents?
- 16:01Why did Arize choose to open source AI observability tooling instead of building a closed SaaS product?
- 19:05What security challenges does Arize face when open sourcing AI agent observability?
- 23:44How should open source maintainers handle pull requests from AI coding agents?