Agno → OpenObserve

Automatically capture agent runs, tool calls, memory lookups, and LLM invocations for every Agno agent in your Python application.

Prerequisites

Python 3.9+
An OpenObserve account (cloud or self-hosted)
Your OpenObserve organisation ID and Base64-encoded auth token
An OpenAI API key (or whichever model backend your agents use)

Installation

pip install openobserve-telemetry-sdk "openinference-instrumentation-agno==0.1.25" "agno==2.2.0" python-dotenv

Configuration

Create a .env file in your project root:

# OpenObserve instance URL
# Default for self-hosted: http://localhost:5080
OPENOBSERVE_URL=https://api.openobserve.ai/

# Your OpenObserve organisation slug or ID
OPENOBSERVE_ORG=your_org_id

# Basic auth token — Base64-encoded "email:password"
OPENOBSERVE_AUTH_TOKEN=Basic <your_base64_token>

# LLM provider key
OPENAI_API_KEY=your-openai-key

Instrumentation

Call AgnoInstrumentor().instrument() before importing Agno.

from dotenv import load_dotenv
load_dotenv()

from openinference.instrumentation.agno import AgnoInstrumentor
from openobserve import openobserve_init

AgnoInstrumentor().instrument()
openobserve_init()

from agno.agent import Agent
from agno.models.openai import OpenAIChat

agent = Agent(
    model=OpenAIChat(id="gpt-4o-mini"),
    instructions="Answer questions concisely in one sentence.",
)

response = agent.run("What is OpenTelemetry?")
print(response.content)

Agent with tools

from agno.tools.duckduckgo import DuckDuckGoTools

agent = Agent(
    model=OpenAIChat(id="gpt-4o-mini"),
    tools=[DuckDuckGoTools()],
    instructions="Use DuckDuckGo to search for current information.",
)

response = agent.run("What is the latest version of Python?")
print(response.content)

What Gets Captured

Each agent.run() call produces a root AGENT span with a child LLM span per model call, and additional TOOL spans when tools are invoked.

LLM span

Attribute	Description
`openinference_span_kind`	`LLM`
`operation_name`	`OpenAIChat.invoke`
`llm_model_name`	Model used (e.g. `gpt-4o-mini`)
`llm_provider`	`OpenAI`
`llm_observation_type`	`GENERATION`
`llm_token_count_prompt`	Input token count
`llm_token_count_completion`	Output token count
`llm_usage_tokens_input`	Input tokens (numeric)
`llm_usage_tokens_output`	Output tokens (numeric)
`llm_usage_tokens_total`	Total tokens consumed
`llm_usage_cost_input`	Estimated input cost in USD
`llm_usage_cost_output`	Estimated output cost in USD
`gen_ai_response_model`	Model that handled the request
`duration`	Span latency
`span_status`	`OK` on success, `ERROR` on failure

Viewing Traces

Log in to OpenObserve and navigate to Traces in the left sidebar
Click any root agent span to open the waterfall view
Expand the trace to see the child OpenAIChat.invoke LLM span with token counts and cost
Filter by llm_model_name to compare performance across different models

Agno trace in OpenObserve

With Agno instrumented, every agent invocation is recorded in OpenObserve with a full span hierarchy. From here you can track token usage per agent run, measure tool latency, and monitor error rates across different agent configurations.