OpenInference instrumentation for Ollama
An instrumentation package for observing calls made through the Ollama Python client.
Published by Arize AI
How to use it
Send a fixed prompt to a local model and compare the request span across repeated runs. Note which runs start with a model that is already loaded.
Run pip install openinference-instrumentation-ollama in the client process and configure a tracing backend. Start the Ollama service locally and test with a model already downloaded to that host.
- Compatibility
- Python, Ollama, OpenTelemetry
- Monitoring areas
- Local model tracing, Request latency
- Access
- Local Python instrumentation. A compatible OpenTelemetry receiver is needed to view traces.