Anannas → OpenObserve

Automatically capture token usage, latency, and model metadata for every Anannas inference call. Anannas is an LLM gateway providing access to 500+ models from OpenAI, Anthropic, Google, and other providers through a single OpenAI-compatible endpoint. Instrumentation uses the standard OpenAI instrumentor pointed at the Anannas base URL.

Prerequisites

Python 3.8+
An OpenObserve account (cloud or self-hosted)
Your OpenObserve organisation ID and Base64-encoded auth token
An Anannas account and API key
A provider API key configured in your Anannas account (e.g. Anthropic, OpenAI)

Installation

pip install openobserve-telemetry-sdk openinference-instrumentation-openai openai python-dotenv

Configuration

Create a .env file in your project root:

OPENOBSERVE_URL=http://localhost:5080/
OPENOBSERVE_ORG=default
OPENOBSERVE_AUTH_TOKEN=Basic <your_base64_token>
ANANNAS_API_KEY=your-anannas-api-key
ANTHROPIC_API_KEY=your-anthropic-api-key

Instrumentation

Call OpenAIInstrumentor().instrument() before creating the client. Pass your Anannas key as the API key and set the base URL to https://anannas.ai/v1. Use default_headers to forward your provider API key via x-provider-api-key.

from dotenv import load_dotenv
load_dotenv()

from openinference.instrumentation.openai import OpenAIInstrumentor
from openobserve import openobserve_init

OpenAIInstrumentor().instrument()
openobserve_init(service_name="anannas")

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["ANANNAS_API_KEY"],
    base_url="https://anannas.ai/v1",
    default_headers={"x-provider-api-key": os.environ["ANTHROPIC_API_KEY"]},
)

response = client.chat.completions.create(
    model="claude-haiku-4.5",
    messages=[{"role": "user", "content": "Explain distributed tracing in one sentence."}],
)
print(response.choices[0].message.content)

What Gets Captured

Attribute	Description
`operation_name`	`ChatCompletion`
`llm_system`	`openai` (OpenAI-compatible client)
`llm_model_name`	Resolved model name (e.g. `claude-haiku-4.5`)
`llm_request_parameters_model`	Model name sent in the request
`llm_observation_type`	`GENERATION`
`llm_token_count_prompt`	Prompt tokens consumed
`llm_token_count_completion`	Completion tokens returned
`llm_token_count_total`	Total tokens consumed
`llm_usage_tokens_input`	Input tokens
`llm_usage_tokens_output`	Output tokens
`openinference_span_kind`	`LLM`
`duration`	End-to-end request latency
`span_status`	`OK` on success, `ERROR` on failure

Viewing Traces

Log in to OpenObserve and navigate to Traces
Filter by operation_name = ChatCompletion to see all inference calls
Use llm_model_name to identify which provider model handled each request
Filter by span_status = ERROR to find failed requests

Anannas trace in OpenObserve

Next Steps

With Anannas instrumented, every inference call is recorded in OpenObserve regardless of which underlying model was used. From here you can compare latency across providers, monitor token usage, and set alerts on error rates.