andre/crewAI
1
0
Fork 0
mirror of https://github.com/crewAIInc/crewAI.git synced 2025-12-15 20:08:29 +00:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

chore: remove AgentOps integration (#3334)
Some checks failed
Notify Downstream / notify-downstream (push) Has been cancelled
Details
Mark stale issues and pull requests / stale (push) Has been cancelled
Details

Browse Source
This commit is contained in:
Greyson LaLonde
2025-08-17 23:07:41 -04:00
committed by GitHub
parent 04a03d332f
commit 947c9552f0
19 changed files with 4 additions and 742 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
.gitignore vendored
View File

@@ -21,7 +21,6 @@ crew_tasks_output.json
.mypy_cache
.ruff_cache
.venv
agentops.log
test_flow.html
crewairules.mdc
plan.md

3
docs/docs.json
View File

@@ -226,7 +226,6 @@
"group": "Observability",
"pages": [
"en/observability/overview",
"en/observability/agentops",
"en/observability/arize-phoenix",
"en/observability/langdb",
"en/observability/langfuse",
@@ -566,7 +565,6 @@
"group": "Observabilidade",
"pages": [
"pt-BR/observability/overview",
"pt-BR/observability/agentops",
"pt-BR/observability/arize-phoenix",
"pt-BR/observability/langdb",
"pt-BR/observability/langfuse",
@@ -914,7 +912,6 @@
"group": "오브저버빌리티",
"pages": [
"ko/observability/overview",
"ko/observability/agentops",
"ko/observability/arize-phoenix",
"ko/observability/langdb",
"ko/observability/langfuse",

80
docs/en/concepts/event-listener.mdx
View File

@@ -177,14 +177,7 @@ class MyCustomCrew:
# Your crew implementation...
```
This is exactly how CrewAI's built-in `agentops_listener` is registered. In the CrewAI codebase, you'll find:
```python
# src/crewai/utilities/events/third_party/__init__.py
from .agentops_listener import agentops_listener
```
This ensures the `agentops_listener` is loaded when the `crewai.utilities.events` package is imported.
This is how third-party event listeners are registered in the CrewAI codebase.
## Available Event Types
@@ -280,77 +273,6 @@ The structure of the event object depends on the event type, but all events inhe
Additional fields vary by event type. For example, `CrewKickoffCompletedEvent` includes `crew_name` and `output` fields.
## Real-World Example: Integration with AgentOps
CrewAI includes an example of a third-party integration with [AgentOps](https://github.com/AgentOps-AI/agentops), a monitoring and observability platform for AI agents. Here's how it's implemented:
```python
from typing import Optional
from crewai.utilities.events import (
CrewKickoffCompletedEvent,
ToolUsageErrorEvent,
ToolUsageStartedEvent,
)
from crewai.utilities.events.base_event_listener import BaseEventListener
from crewai.utilities.events.crew_events import CrewKickoffStartedEvent
from crewai.utilities.events.task_events import TaskEvaluationEvent
try:
import agentops
AGENTOPS_INSTALLED = True
except ImportError:
AGENTOPS_INSTALLED = False
class AgentOpsListener(BaseEventListener):
tool_event: Optional["agentops.ToolEvent"] = None
session: Optional["agentops.Session"] = None
def __init__(self):
super().__init__()
def setup_listeners(self, crewai_event_bus):
if not AGENTOPS_INSTALLED:
return
@crewai_event_bus.on(CrewKickoffStartedEvent)
def on_crew_kickoff_started(source, event: CrewKickoffStartedEvent):
self.session = agentops.init()
for agent in source.agents:
if self.session:
self.session.create_agent(
name=agent.role,
agent_id=str(agent.id),
)
@crewai_event_bus.on(CrewKickoffCompletedEvent)
def on_crew_kickoff_completed(source, event: CrewKickoffCompletedEvent):
if self.session:
self.session.end_session(
end_state="Success",
end_state_reason="Finished Execution",
)
@crewai_event_bus.on(ToolUsageStartedEvent)
def on_tool_usage_started(source, event: ToolUsageStartedEvent):
self.tool_event = agentops.ToolEvent(name=event.tool_name)
if self.session:
self.session.record(self.tool_event)
@crewai_event_bus.on(ToolUsageErrorEvent)
def on_tool_usage_error(source, event: ToolUsageErrorEvent):
agentops.ErrorEvent(exception=event.error, trigger_event=self.tool_event)
```
This listener initializes an AgentOps session when a Crew starts, registers agents with AgentOps, tracks tool usage, and ends the session when the Crew completes.
The AgentOps listener is registered in CrewAI's event system through the import in `src/crewai/utilities/events/third_party/__init__.py`:
```python
from .agentops_listener import agentops_listener
```
This ensures the `agentops_listener` is loaded when the `crewai.utilities.events` package is imported.
## Advanced Usage: Scoped Handlers

126
docs/en/observability/agentops.mdx
View File

@@ -1,126 +0,0 @@
---
title: AgentOps Integration
description: Understanding and logging your agent performance with AgentOps.
icon: paperclip
---
# Introduction
Observability is a key aspect of developing and deploying conversational AI agents. It allows developers to understand how their agents are performing,
how their agents are interacting with users, and how their agents use external tools and APIs.
AgentOps is a product independent of CrewAI that provides a comprehensive observability solution for agents.
## AgentOps
[AgentOps](https://agentops.ai/?=crew) provides session replays, metrics, and monitoring for agents.
At a high level, AgentOps gives you the ability to monitor cost, token usage, latency, agent failures, session-wide statistics, and more.
For more info, check out the [AgentOps Repo](https://github.com/AgentOps-AI/agentops).
### Overview
AgentOps provides monitoring for agents in development and production.
It provides a dashboard for tracking agent performance, session replays, and custom reporting.
Additionally, AgentOps provides session drilldowns for viewing Crew agent interactions, LLM calls, and tool usage in real-time.
This feature is useful for debugging and understanding how agents interact with users as well as other agents.
![Overview of a select series of agent session runs](/images/agentops-overview.png)
![Overview of session drilldowns for examining agent runs](/images/agentops-session.png)
![Viewing a step-by-step agent replay execution graph](/images/agentops-replay.png)
### Features
- **LLM Cost Management and Tracking**: Track spend with foundation model providers.
- **Replay Analytics**: Watch step-by-step agent execution graphs.
- **Recursive Thought Detection**: Identify when agents fall into infinite loops.
- **Custom Reporting**: Create custom analytics on agent performance.
- **Analytics Dashboard**: Monitor high-level statistics about agents in development and production.
- **Public Model Testing**: Test your agents against benchmarks and leaderboards.
- **Custom Tests**: Run your agents against domain-specific tests.
- **Time Travel Debugging**: Restart your sessions from checkpoints.
- **Compliance and Security**: Create audit logs and detect potential threats such as profanity and PII leaks.
- **Prompt Injection Detection**: Identify potential code injection and secret leaks.
### Using AgentOps
<Steps>
<Step title="Create an API Key">
Create a user API key here: [Create API Key](https://app.agentops.ai/account)
</Step>
<Step title="Configure Your Environment">
Add your API key to your environment variables:
```bash
AGENTOPS_API_KEY=<YOUR_AGENTOPS_API_KEY>
```
</Step>
<Step title="Install AgentOps">
Install AgentOps with:
```bash
pip install 'crewai[agentops]'
```
or
```bash
pip install agentops
```
</Step>
<Step title="Initialize AgentOps">
Before using `Crew` in your script, include these lines:
```python
import agentops
agentops.init()
```
This will initiate an AgentOps session as well as automatically track Crew agents. For further info on how to outfit more complex agentic systems,
check out the [AgentOps documentation](https://docs.agentops.ai) or join the [Discord](https://discord.gg/j4f3KbeH).
</Step>
</Steps>
### Crew + AgentOps Examples
<CardGroup cols={3}>
<Card
title="Job Posting"
color="#F3A78B"
href="https://github.com/joaomdmoura/crewAI-examples/tree/main/job-posting"
icon="briefcase"
iconType="solid"
>
Example of a Crew agent that generates job posts.
</Card>
<Card
title="Markdown Validator"
color="#F3A78B"
href="https://github.com/joaomdmoura/crewAI-examples/tree/main/markdown_validator"
icon="markdown"
iconType="solid"
>
Example of a Crew agent that validates Markdown files.
</Card>
<Card
title="Instagram Post"
color="#F3A78B"
href="https://github.com/joaomdmoura/crewAI-examples/tree/main/instagram_post"
icon="square-instagram"
iconType="brands"
>
Example of a Crew agent that generates Instagram posts.
</Card>
</CardGroup>
### Further Information
To get started, create an [AgentOps account](https://agentops.ai/?=crew).
For feature requests or bug reports, please reach out to the AgentOps team on the [AgentOps Repo](https://github.com/AgentOps-AI/agentops).
#### Extra links
<a href="https://twitter.com/agentopsai/">🐦 Twitter</a>
<span>&nbsp;&nbsp;•&nbsp;&nbsp;</span>
<a href="https://discord.gg/JHPt4C7r">📢 Discord</a>
<span>&nbsp;&nbsp;•&nbsp;&nbsp;</span>
<a href="https://app.agentops.ai/?=crew">🖇️ AgentOps Dashboard</a>
<span>&nbsp;&nbsp;•&nbsp;&nbsp;</span>
<a href="https://docs.agentops.ai/introduction">📙 Documentation</a>

3
docs/en/observability/overview.mdx
View File

@@ -21,9 +21,6 @@ Observability is crucial for understanding how your CrewAI agents perform, ident
### Monitoring & Tracing Platforms
<CardGroup cols={2}>
<Card title="AgentOps" icon="paperclip" href="/en/observability/agentops">
Session replays, metrics, and monitoring for agent development and production.
</Card>
<Card title="LangDB" icon="database" href="/en/observability/langdb">
End-to-end tracing for CrewAI workflows with automatic agent interaction capture.

BIN
docs/images/agentops-overview.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 288 KiB

BIN
docs/images/agentops-replay.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 419 KiB

BIN
docs/images/agentops-session.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 263 KiB

80
docs/ko/concepts/event-listener.mdx
View File

@@ -177,14 +177,7 @@ class MyCustomCrew:
# Your crew implementation...
```
이것이 바로 CrewAI의 내장 `agentops_listener`가 등록되는 방식과 동일합니다. CrewAI 코드베이스에서는 다음과 같이 되어 있습니다:
```python
# src/crewai/utilities/events/third_party/__init__.py
from .agentops_listener import agentops_listener
```
이렇게 하면 `crewai.utilities.events` 패키지가 임포트될 때 `agentops_listener`가 자동으로 로드됩니다.
이것이 CrewAI 코드베이스에서 서드파티 이벤트 리스너가 등록되는 방식입니다.
## 사용 가능한 이벤트 유형
@@ -280,77 +273,6 @@ CrewAI는 여러분이 청취할 수 있는 다양한 이벤트를 제공합니
추가 필드는 이벤트 타입에 따라 다릅니다. 예를 들어, `CrewKickoffCompletedEvent`에는 `crew_name`과 `output` 필드가 포함됩니다.
## 실제 예시: AgentOps와의 통합
CrewAI는 AI 에이전트를 위한 모니터링 및 관찰 플랫폼인 [AgentOps](https://github.com/AgentOps-AI/agentops)와의 서드파티 통합 예시를 포함하고 있습니다. 구현 방식은 다음과 같습니다:
```python
from typing import Optional
from crewai.utilities.events import (
CrewKickoffCompletedEvent,
ToolUsageErrorEvent,
ToolUsageStartedEvent,
)
from crewai.utilities.events.base_event_listener import BaseEventListener
from crewai.utilities.events.crew_events import CrewKickoffStartedEvent
from crewai.utilities.events.task_events import TaskEvaluationEvent
try:
import agentops
AGENTOPS_INSTALLED = True
except ImportError:
AGENTOPS_INSTALLED = False
class AgentOpsListener(BaseEventListener):
tool_event: Optional["agentops.ToolEvent"] = None
session: Optional["agentops.Session"] = None
def __init__(self):
super().__init__()
def setup_listeners(self, crewai_event_bus):
if not AGENTOPS_INSTALLED:
return
@crewai_event_bus.on(CrewKickoffStartedEvent)
def on_crew_kickoff_started(source, event: CrewKickoffStartedEvent):
self.session = agentops.init()
for agent in source.agents:
if self.session:
self.session.create_agent(
name=agent.role,
agent_id=str(agent.id),
)
@crewai_event_bus.on(CrewKickoffCompletedEvent)
def on_crew_kickoff_completed(source, event: CrewKickoffCompletedEvent):
if self.session:
self.session.end_session(
end_state="Success",
end_state_reason="Finished Execution",
)
@crewai_event_bus.on(ToolUsageStartedEvent)
def on_tool_usage_started(source, event: ToolUsageStartedEvent):
self.tool_event = agentops.ToolEvent(name=event.tool_name)
if self.session:
self.session.record(self.tool_event)
@crewai_event_bus.on(ToolUsageErrorEvent)
def on_tool_usage_error(source, event: ToolUsageErrorEvent):
agentops.ErrorEvent(exception=event.error, trigger_event=self.tool_event)
```
이 listener는 crew가 시작될 때 AgentOps 세션을 초기화하고, agent를 AgentOps에 등록하며, 도구 사용을 추적하고, crew가 완료되면 세션을 종료합니다.
AgentOps listener는 `src/crewai/utilities/events/third_party/__init__.py` 파일의 import를 통해 CrewAI 이벤트 시스템에 등록됩니다:
```python
from .agentops_listener import agentops_listener
```
이렇게 하면 `crewai.utilities.events` 패키지가 import될 때 `agentops_listener`가 로드되는 것이 보장됩니다.
## 고급 사용법: Scoped Handlers

124
docs/ko/observability/agentops.mdx
View File

@@ -1,124 +0,0 @@
---
title: AgentOps 통합
description: AgentOps를 사용하여 에이전트 성능을 이해하고 로깅하기
icon: paperclip
---
# 소개
Observability는 대화형 AI 에이전트를 개발하고 배포하는 데 있어 핵심적인 요소입니다. 이는 개발자가 에이전트의 성능을 이해하고, 에이전트가 사용자와 어떻게 상호작용하는지, 그리고 에이전트가 외부 도구와 API를 어떻게 사용하는지를 파악할 수 있게 해줍니다.
AgentOps는 CrewAI와 독립적인 제품으로, 에이전트를 위한 종합적인 observability 솔루션을 제공합니다.
## AgentOps
[AgentOps](https://agentops.ai/?=crew)은 에이전트에 대한 세션 리플레이, 메트릭, 모니터링을 제공합니다.
AgentOps는 높은 수준에서 비용, 토큰 사용량, 대기 시간, 에이전트 실패, 세션 전체 통계 등 다양한 항목을 모니터링할 수 있는 기능을 제공합니다.
더 자세한 내용은 [AgentOps Repo](https://github.com/AgentOps-AI/agentops)를 확인하세요.
### 개요
AgentOps는 개발 및 프로덕션 환경에서 에이전트에 대한 모니터링을 제공합니다.
에이전트 성능, 세션 리플레이, 맞춤형 리포팅을 추적할 수 있는 대시보드를 제공합니다.
또한, AgentOps는 Crew 에이전트 상호작용, LLM 호출, 툴 사용을 실시간으로 볼 수 있는 세션 드릴다운 기능을 제공합니다.
이 기능은 에이전트가 사용자 및 다른 에이전트와 어떻게 상호작용하는지 디버깅하고 이해하는 데 유용합니다.
![선택된 에이전트 세션 실행 시리즈의 개요](/images/agentops-overview.png)
![에이전트 실행을 조사하기 위한 세션 드릴다운 개요](/images/agentops-session.png)
![단계별 에이전트 리플레이 실행 그래프 보기](/images/agentops-replay.png)
### 특징
- **LLM 비용 관리 및 추적**: 기반 모델 공급자와의 지출을 추적합니다.
- **재생 분석**: 단계별 에이전트 실행 그래프를 시청할 수 있습니다.
- **재귀적 사고 감지**: 에이전트가 무한 루프에 빠졌는지 식별합니다.
- **맞춤형 보고서**: 에이전트 성능에 대한 맞춤형 분석을 생성합니다.
- **분석 대시보드**: 개발 및 운영 중인 에이전트에 대한 상위 수준 통계를 모니터링합니다.
- **공개 모델 테스트**: 벤치마크 및 리더보드를 통해 에이전트를 테스트할 수 있습니다.
- **맞춤형 테스트**: 도메인별 테스트로 에이전트를 실행합니다.
- **타임 트래블 디버깅**: 체크포인트에서 세션을 재시작합니다.
- **컴플라이언스 및 보안**: 감사 로그를 생성하고 욕설 및 PII 유출과 같은 잠재적 위협을 감지합니다.
- **프롬프트 인젝션 감지**: 잠재적 코드 인젝션 및 시크릿 유출을 식별합니다.
### AgentOps 사용하기
<Steps>
<Step title="API 키 생성">
사용자 API 키를 여기서 생성하세요: [API 키 생성](https://app.agentops.ai/account)
</Step>
<Step title="환경 설정">
API 키를 환경 변수에 추가하세요:
```bash
AGENTOPS_API_KEY=<YOUR_AGENTOPS_API_KEY>
```
</Step>
<Step title="AgentOps 설치">
다음 명령어로 AgentOps를 설치하세요:
```bash
pip install 'crewai[agentops]'
```
또는
```bash
pip install agentops
```
</Step>
<Step title="AgentOps 초기화">
스크립트에서 `Crew`를 사용하기 전에 다음 코드를 포함하세요:
```python
import agentops
agentops.init()
```
이렇게 하면 AgentOps 세션이 시작되고 Crew 에이전트가 자동으로 추적됩니다. 더 복잡한 agentic 시스템을 구성하는 방법에 대한 자세한 정보는 [AgentOps 문서](https://docs.agentops.ai) 또는 [Discord](https://discord.gg/j4f3KbeH)를 참조하세요.
</Step>
</Steps>
### Crew + AgentOps 예시
<CardGroup cols={3}>
<Card
title="Job Posting"
color="#F3A78B"
href="https://github.com/joaomdmoura/crewAI-examples/tree/main/job-posting"
icon="briefcase"
iconType="solid"
>
채용 공고를 생성하는 Crew agent의 예시입니다.
</Card>
<Card
title="Markdown Validator"
color="#F3A78B"
href="https://github.com/joaomdmoura/crewAI-examples/tree/main/markdown_validator"
icon="markdown"
iconType="solid"
>
Markdown 파일을 검증하는 Crew agent의 예시입니다.
</Card>
<Card
title="Instagram Post"
color="#F3A78B"
href="https://github.com/joaomdmoura/crewAI-examples/tree/main/instagram_post"
icon="square-instagram"
iconType="brands"
>
Instagram 게시물을 생성하는 Crew agent의 예시입니다.
</Card>
</CardGroup>
### 추가 정보
시작하려면 [AgentOps 계정](https://agentops.ai/?=crew)을 생성하세요.
기능 요청이나 버그 보고가 필요하시면 [AgentOps Repo](https://github.com/AgentOps-AI/agentops)에서 AgentOps 팀에 문의해 주세요.
#### 추가 링크
<a href="https://twitter.com/agentopsai/">🐦 트위터</a>
<span>&nbsp;&nbsp;•&nbsp;&nbsp;</span>
<a href="https://discord.gg/JHPt4C7r">📢 디스코드</a>
<span>&nbsp;&nbsp;•&nbsp;&nbsp;</span>
<a href="https://app.agentops.ai/?=crew">🖇️ AgentOps 대시보드</a>
<span>&nbsp;&nbsp;•&nbsp;&nbsp;</span>
<a href="https://docs.agentops.ai/introduction">📙 문서화</a>

3
docs/ko/observability/overview.mdx
View File

@@ -21,9 +21,6 @@ icon: "face-smile"
### 모니터링 & 트레이싱 플랫폼
<CardGroup cols={2}>
<Card title="AgentOps" icon="paperclip" href="/ko/observability/agentops">
에이전트 개발 및 운영을 위한 세션 리플레이, 메트릭, 모니터링 제공.
</Card>
<Card title="LangDB" icon="database" href="/ko/observability/langdb">
자동 에이전트 상호작용 캡처를 포함한 CrewAI 워크플로의 엔드-투-엔드 트레이싱.

80
docs/pt-BR/concepts/event-listener.mdx
View File

@@ -177,14 +177,7 @@ class MyCustomCrew:
# Sua implementação do crew...
```
É exatamente assim que o `agentops_listener` integrado do CrewAI é registrado. No código-fonte do CrewAI, você encontrará:
```python
# src/crewai/utilities/events/third_party/__init__.py
from .agentops_listener import agentops_listener
```
Isso garante que o `agentops_listener` seja carregado quando o pacote `crewai.utilities.events` for importado.
É assim que listeners de eventos de terceiros são registrados no código do CrewAI.
## Tipos de Eventos Disponíveis
@@ -269,77 +262,6 @@ A estrutura do objeto de evento depende do tipo do evento, mas todos herdam de `
Campos adicionais variam pelo tipo de evento. Por exemplo, `CrewKickoffCompletedEvent` inclui os campos `crew_name` e `output`.
## Exemplo Real: Integração com AgentOps
O CrewAI inclui um exemplo de integração com [AgentOps](https://github.com/AgentOps-AI/agentops), uma plataforma de monitoramento e observabilidade para agentes de IA. Veja como é implementado:
```python
from typing import Optional
from crewai.utilities.events import (
CrewKickoffCompletedEvent,
ToolUsageErrorEvent,
ToolUsageStartedEvent,
)
from crewai.utilities.events.base_event_listener import BaseEventListener
from crewai.utilities.events.crew_events import CrewKickoffStartedEvent
from crewai.utilities.events.task_events import TaskEvaluationEvent
try:
import agentops
AGENTOPS_INSTALLED = True
except ImportError:
AGENTOPS_INSTALLED = False
class AgentOpsListener(BaseEventListener):
tool_event: Optional["agentops.ToolEvent"] = None
session: Optional["agentops.Session"] = None
def __init__(self):
super().__init__()
def setup_listeners(self, crewai_event_bus):
if not AGENTOPS_INSTALLED:
return
@crewai_event_bus.on(CrewKickoffStartedEvent)
def on_crew_kickoff_started(source, event: CrewKickoffStartedEvent):
self.session = agentops.init()
for agent in source.agents:
if self.session:
self.session.create_agent(
name=agent.role,
agent_id=str(agent.id),
)
@crewai_event_bus.on(CrewKickoffCompletedEvent)
def on_crew_kickoff_completed(source, event: CrewKickoffCompletedEvent):
if self.session:
self.session.end_session(
end_state="Success",
end_state_reason="Finished Execution",
)
@crewai_event_bus.on(ToolUsageStartedEvent)
def on_tool_usage_started(source, event: ToolUsageStartedEvent):
self.tool_event = agentops.ToolEvent(name=event.tool_name)
if self.session:
self.session.record(self.tool_event)
@crewai_event_bus.on(ToolUsageErrorEvent)
def on_tool_usage_error(source, event: ToolUsageErrorEvent):
agentops.ErrorEvent(exception=event.error, trigger_event=self.tool_event)
```
Esse listener inicializa uma sessão do AgentOps quando um Crew inicia, cadastra agentes no AgentOps, rastreia o uso de ferramentas e finaliza a sessão quando o Crew é concluído.
O listener AgentOps é registrado no sistema de eventos do CrewAI via importação em `src/crewai/utilities/events/third_party/__init__.py`:
```python
from .agentops_listener import agentops_listener
```
Isso garante que o `agentops_listener` seja carregado quando o pacote `crewai.utilities.events` for importado.
## Uso Avançado: Handlers Escopados

126
docs/pt-BR/observability/agentops.mdx
View File

@@ -1,126 +0,0 @@
---
title: Integração com AgentOps
description: Entendendo e registrando a performance do seu agente com AgentOps.
icon: paperclip
---
# Introdução
Observabilidade é um aspecto fundamental no desenvolvimento e implantação de agentes de IA conversacional. Ela permite que desenvolvedores compreendam como seus agentes estão performando,
como eles estão interagindo com os usuários e como utilizam ferramentas externas e APIs.
AgentOps é um produto independente do CrewAI que fornece uma solução completa de observabilidade para agentes.
## AgentOps
[AgentOps](https://agentops.ai/?=crew) oferece replay de sessões, métricas e monitoramento para agentes.
Em um alto nível, o AgentOps oferece a capacidade de monitorar custos, uso de tokens, latência, falhas do agente, estatísticas de sessão e muito mais.
Para mais informações, confira o [Repositório do AgentOps](https://github.com/AgentOps-AI/agentops).
### Visão Geral
AgentOps fornece monitoramento para agentes em desenvolvimento e produção.
Disponibiliza um dashboard para acompanhamento de performance dos agentes, replay de sessões e relatórios personalizados.
Além disso, o AgentOps traz análises detalhadas das sessões para visualizar interações do agente Crew, chamadas LLM e uso de ferramentas em tempo real.
Esse recurso é útil para depuração e entendimento de como os agentes interagem com usuários e entre si.
![Visão geral de uma série selecionada de execuções de sessões do agente](/images/agentops-overview.png)
![Visão geral das análises detalhadas de sessões para examinar execuções de agentes](/images/agentops-session.png)
![Visualizando um gráfico de execução passo a passo do replay do agente](/images/agentops-replay.png)
### Funcionalidades
- **Gerenciamento e Rastreamento de Custos de LLM**: Acompanhe gastos com provedores de modelos fundamentais.
- **Análises de Replay**: Assista gráficos de execução do agente, passo a passo.
- **Detecção de Pensamento Recursivo**: Identifique quando agentes entram em loops infinitos.
- **Relatórios Personalizados**: Crie análises customizadas sobre a performance dos agentes.
- **Dashboard Analítico**: Monitore estatísticas gerais de agentes em desenvolvimento e produção.
- **Teste de Modelos Públicos**: Teste seus agentes em benchmarks e rankings.
- **Testes Personalizados**: Execute seus agentes em testes específicos de domínio.
- **Depuração com Viagem no Tempo**: Reinicie suas sessões a partir de checkpoints.
- **Conformidade e Segurança**: Crie registros de auditoria e detecte possíveis ameaças como uso de palavrões e vazamento de dados pessoais.
- **Detecção de Prompt Injection**: Identifique possíveis injeções de código e vazamentos de segredos.
### Utilizando o AgentOps
<Steps>
<Step title="Crie uma Chave de API">
Crie uma chave de API de usuário aqui: [Create API Key](https://app.agentops.ai/account)
</Step>
<Step title="Configure seu Ambiente">
Adicione sua chave API nas variáveis de ambiente:
```bash
AGENTOPS_API_KEY=<YOUR_AGENTOPS_API_KEY>
```
</Step>
<Step title="Instale o AgentOps">
Instale o AgentOps com:
```bash
pip install 'crewai[agentops]'
```
ou
```bash
pip install agentops
```
</Step>
<Step title="Inicialize o AgentOps">
Antes de utilizar o `Crew` no seu script, inclua estas linhas:
```python
import agentops
agentops.init()
```
Isso irá iniciar uma sessão do AgentOps e também rastrear automaticamente os agentes Crew. Para mais detalhes sobre como adaptar sistemas de agentes mais complexos,
confira a [documentação do AgentOps](https://docs.agentops.ai) ou participe do [Discord](https://discord.gg/j4f3KbeH).
</Step>
</Steps>
### Exemplos de Crew + AgentOps
<CardGroup cols={3}>
<Card
title="Vaga de Emprego"
color="#F3A78B"
href="https://github.com/joaomdmoura/crewAI-examples/tree/main/job-posting"
icon="briefcase"
iconType="solid"
>
Exemplo de um agente Crew que gera vagas de emprego.
</Card>
<Card
title="Validador de Markdown"
color="#F3A78B"
href="https://github.com/joaomdmoura/crewAI-examples/tree/main/markdown_validator"
icon="markdown"
iconType="solid"
>
Exemplo de um agente Crew que valida arquivos Markdown.
</Card>
<Card
title="Post no Instagram"
color="#F3A78B"
href="https://github.com/joaomdmoura/crewAI-examples/tree/main/instagram_post"
icon="square-instagram"
iconType="brands"
>
Exemplo de um agente Crew que gera posts para Instagram.
</Card>
</CardGroup>
### Mais Informações
Para começar, crie uma [conta AgentOps](https://agentops.ai/?=crew).
Para sugestões de funcionalidades ou relatos de bugs, entre em contato com o time do AgentOps pelo [Repositório do AgentOps](https://github.com/AgentOps-AI/agentops).
#### Links Extras
<a href="https://twitter.com/agentopsai/">🐦 Twitter</a>
<span>&nbsp;&nbsp;•&nbsp;&nbsp;</span>
<a href="https://discord.gg/JHPt4C7r">📢 Discord</a>
<span>&nbsp;&nbsp;•&nbsp;&nbsp;</span>
<a href="https://app.agentops.ai/?=crew">🖇️ Dashboard AgentOps</a>
<span>&nbsp;&nbsp;•&nbsp;&nbsp;</span>
<a href="https://docs.agentops.ai/introduction">📙 Documentação</a>

3
docs/pt-BR/observability/overview.mdx
View File

@@ -21,9 +21,6 @@ A observabilidade é fundamental para entender como seus agentes CrewAI estão d
### Plataformas de Monitoramento e Rastreamento
<CardGroup cols={2}>
<Card title="AgentOps" icon="paperclip" href="/pt-BR/observability/agentops">
Replays de sessões, métricas e monitoramento para desenvolvimento e produção de agentes.
</Card>
<Card title="LangDB" icon="database" href="/pt-BR/observability/langdb">
Rastreamento ponta a ponta para fluxos de trabalho CrewAI com captura automática de interações de agentes.

1
pyproject.toml
View File

@@ -52,7 +52,6 @@ tools = ["crewai-tools~=0.62.0"]
embeddings = [
"tiktoken~=0.8.0"
]
agentops = ["agentops==0.3.18"]
pdfplumber = [
"pdfplumber>=0.11.4",
]

3
src/crewai/utilities/events/__init__.py
View File

@@ -67,11 +67,9 @@ from .memory_events import (
# events
from .event_listener import EventListener
from .third_party.agentops_listener import agentops_listener
__all__ = [
"EventListener",
"agentops_listener",
"CrewAIEventsBus",
"crewai_event_bus",
"AgentExecutionStartedEvent",
@@ -105,7 +103,6 @@ __all__ = [
"MemoryRetrievalStartedEvent",
"MemoryRetrievalCompletedEvent",
"EventListener",
"agentops_listener",
"CrewKickoffStartedEvent",
"CrewKickoffCompletedEvent",
"CrewKickoffFailedEvent",

1
src/crewai/utilities/events/third_party/__init__.py vendored
View File

@@ -1 +0,0 @@
from .agentops_listener import agentops_listener

67
src/crewai/utilities/events/third_party/agentops_listener.py vendored
View File

@@ -1,67 +0,0 @@
from typing import Optional
from crewai.utilities.events import (
CrewKickoffCompletedEvent,
ToolUsageErrorEvent,
ToolUsageStartedEvent,
)
from crewai.utilities.events.base_event_listener import BaseEventListener
from crewai.utilities.events.crew_events import CrewKickoffStartedEvent
from crewai.utilities.events.task_events import TaskEvaluationEvent
try:
import agentops
AGENTOPS_INSTALLED = True
except ImportError:
AGENTOPS_INSTALLED = False
class AgentOpsListener(BaseEventListener):
tool_event: Optional["agentops.ToolEvent"] = None
session: Optional["agentops.Session"] = None
def __init__(self):
super().__init__()
def setup_listeners(self, crewai_event_bus):
if not AGENTOPS_INSTALLED:
return
@crewai_event_bus.on(CrewKickoffStartedEvent)
def on_crew_kickoff_started(source, event: CrewKickoffStartedEvent):
self.session = agentops.init()
for agent in source.agents:
if self.session:
self.session.create_agent(
name=agent.role,
agent_id=str(agent.id),
)
@crewai_event_bus.on(CrewKickoffCompletedEvent)
def on_crew_kickoff_completed(source, event: CrewKickoffCompletedEvent):
if self.session:
self.session.end_session(
end_state="Success",
end_state_reason="Finished Execution",
)
@crewai_event_bus.on(ToolUsageStartedEvent)
def on_tool_usage_started(source, event: ToolUsageStartedEvent):
self.tool_event = agentops.ToolEvent(name=event.tool_name)
if self.session:
self.session.record(self.tool_event)
@crewai_event_bus.on(ToolUsageErrorEvent)
def on_tool_usage_error(source, event: ToolUsageErrorEvent):
agentops.ErrorEvent(exception=event.error, trigger_event=self.tool_event)
@crewai_event_bus.on(TaskEvaluationEvent)
def on_task_evaluation(source, event: TaskEvaluationEvent):
if self.session:
self.session.create_agent(
name="Task Evaluator", agent_id=str(source.original_agent.id)
)
agentops_listener = AgentOpsListener()

45
uv.lock generated
View File

@@ -34,22 +34,6 @@ resolution-markers = [
"(python_full_version < '3.11' and platform_machine != 'aarch64' and platform_python_implementation == 'PyPy' and sys_platform == 'linux') or (python_full_version < '3.11' and platform_python_implementation == 'PyPy' and sys_platform != 'darwin' and sys_platform != 'linux')",
]
[[package]]
name = "agentops"
version = "0.3.18"
source = { registry = "https://pypi.org/simple" }
dependencies = [
{ name = "packaging" },
{ name = "psutil" },
{ name = "pyyaml" },
{ name = "requests" },
{ name = "termcolor" },
]
sdist = { url = "https://files.pythonhosted.org/packages/c5/52/46bb2f29b9e5f2e1d8b124296b7794934a9048de635d9e7d6a95e791ad7b/agentops-0.3.18.tar.gz", hash = "sha256:4d509754df7be52579597cc9f53939c5218131a0379463e0ff6f6f40cde9fcc4", size = 55394, upload-time = "2024-11-19T19:06:21.306Z" }
wheels = [
{ url = "https://files.pythonhosted.org/packages/97/8d/bd4cad95dad722dc2d3e4179feab1058ef846828c0e15e51e8bfaea373ee/agentops-0.3.18-py3-none-any.whl", hash = "sha256:8b48d8a1662f276653430fd541c77fa4f9a15a43e881b518ff88ea56925afcf7", size = 58032, upload-time = "2024-11-19T19:06:19.068Z" },
]
[[package]]
name = "aiohappyeyeballs"
version = "2.6.1"
@@ -744,9 +728,6 @@ dependencies = [
]
[package.optional-dependencies]
agentops = [
{ name = "agentops" },
]
aisuite = [
{ name = "aisuite" },
]
@@ -792,7 +773,6 @@ dev = [
[package.metadata]
requires-dist = [
{ name = "agentops", marker = "extra == 'agentops'", specifier = "==0.3.18" },
{ name = "aisuite", marker = "extra == 'aisuite'", specifier = ">=0.1.10" },
{ name = "appdirs", specifier = ">=1.4.4" },
{ name = "blinker", specifier = ">=1.9.0" },
@@ -828,7 +808,7 @@ requires-dist = [
{ name = "tomli-w", specifier = ">=1.1.0" },
{ name = "uv", specifier = ">=0.4.25" },
]
provides-extras = ["agentops", "aisuite", "docling", "embeddings", "mem0", "openpyxl", "pandas", "pdfplumber", "tools"]
provides-extras = ["aisuite", "docling", "embeddings", "mem0", "openpyxl", "pandas", "pdfplumber", "tools"]
[package.metadata.requires-dev]
dev = [
@@ -4013,20 +3993,6 @@ wheels = [
{ url = "https://files.pythonhosted.org/packages/7e/cc/7e77861000a0691aeea8f4566e5d3aa716f2b1dece4a24439437e41d3d25/protobuf-5.29.5-py3-none-any.whl", hash = "sha256:6cf42630262c59b2d8de33954443d94b746c952b01434fc58a417fdbd2e84bd5", size = 172823, upload-time = "2025-05-28T23:51:58.157Z" },
]
[[package]]
name = "psutil"
version = "5.9.8"
source = { registry = "https://pypi.org/simple" }
sdist = { url = "https://files.pythonhosted.org/packages/90/c7/6dc0a455d111f68ee43f27793971cf03fe29b6ef972042549db29eec39a2/psutil-5.9.8.tar.gz", hash = "sha256:6be126e3225486dff286a8fb9a06246a5253f4c7c53b475ea5f5ac934e64194c", size = 503247, upload-time = "2024-01-19T20:47:09.517Z" }
wheels = [
{ url = "https://files.pythonhosted.org/packages/e7/e3/07ae864a636d70a8a6f58da27cb1179192f1140d5d1da10886ade9405797/psutil-5.9.8-cp36-abi3-macosx_10_9_x86_64.whl", hash = "sha256:aee678c8720623dc456fa20659af736241f575d79429a0e5e9cf88ae0605cc81", size = 248702, upload-time = "2024-01-19T20:47:36.303Z" },
{ url = "https://files.pythonhosted.org/packages/b3/bd/28c5f553667116b2598b9cc55908ec435cb7f77a34f2bff3e3ca765b0f78/psutil-5.9.8-cp36-abi3-manylinux_2_12_i686.manylinux2010_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:8cb6403ce6d8e047495a701dc7c5bd788add903f8986d523e3e20b98b733e421", size = 285242, upload-time = "2024-01-19T20:47:39.65Z" },
{ url = "https://files.pythonhosted.org/packages/c5/4f/0e22aaa246f96d6ac87fe5ebb9c5a693fbe8877f537a1022527c47ca43c5/psutil-5.9.8-cp36-abi3-manylinux_2_12_x86_64.manylinux2010_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d06016f7f8625a1825ba3732081d77c94589dca78b7a3fc072194851e88461a4", size = 288191, upload-time = "2024-01-19T20:47:43.078Z" },
{ url = "https://files.pythonhosted.org/packages/6e/f5/2aa3a4acdc1e5940b59d421742356f133185667dd190b166dbcfcf5d7b43/psutil-5.9.8-cp37-abi3-win32.whl", hash = "sha256:bc56c2a1b0d15aa3eaa5a60c9f3f8e3e565303b465dbf57a1b730e7a2b9844e0", size = 251252, upload-time = "2024-01-19T20:47:52.88Z" },
{ url = "https://files.pythonhosted.org/packages/93/52/3e39d26feae7df0aa0fd510b14012c3678b36ed068f7d78b8d8784d61f0e/psutil-5.9.8-cp37-abi3-win_amd64.whl", hash = "sha256:8db4c1b57507eef143a15a6884ca10f7c73876cdf5d51e713151c1236a0e68cf", size = 255090, upload-time = "2024-01-19T20:47:56.019Z" },
{ url = "https://files.pythonhosted.org/packages/05/33/2d74d588408caedd065c2497bdb5ef83ce6082db01289a1e1147f6639802/psutil-5.9.8-cp38-abi3-macosx_11_0_arm64.whl", hash = "sha256:d16bbddf0693323b8c6123dd804100241da461e41d6e332fb0ba6058f630f8c8", size = 249898, upload-time = "2024-01-19T20:47:59.238Z" },
]
[[package]]
name = "ptyprocess"
version = "0.7.0"
@@ -5502,15 +5468,6 @@ wheels = [
{ url = "https://files.pythonhosted.org/packages/e5/30/643397144bfbfec6f6ef821f36f33e57d35946c44a2352d3c9f0ae847619/tenacity-9.1.2-py3-none-any.whl", hash = "sha256:f77bf36710d8b73a50b2dd155c97b870017ad21afe6ab300326b0371b3b05138", size = 28248, upload-time = "2025-04-02T08:25:07.678Z" },
]
[[package]]
name = "termcolor"
version = "2.4.0"
source = { registry = "https://pypi.org/simple" }
sdist = { url = "https://files.pythonhosted.org/packages/10/56/d7d66a84f96d804155f6ff2873d065368b25a07222a6fd51c4f24ef6d764/termcolor-2.4.0.tar.gz", hash = "sha256:aab9e56047c8ac41ed798fa36d892a37aca6b3e9159f3e0c24bc64a9b3ac7b7a", size = 12664, upload-time = "2023-12-01T11:04:51.66Z" }
wheels = [
{ url = "https://files.pythonhosted.org/packages/d9/5f/8c716e47b3a50cbd7c146f45881e11d9414def768b7cd9c5e6650ec2a80a/termcolor-2.4.0-py3-none-any.whl", hash = "sha256:9297c0df9c99445c2412e832e882a7884038a25617c60cea2ad69488d4040d63", size = 7719, upload-time = "2023-12-01T11:04:50.019Z" },
]
[[package]]
name = "tifffile"
version = "2025.5.10"