andre/crewAI
1
0
Fork 0
mirror of https://github.com/crewAIInc/crewAI.git synced 2026-01-07 15:18:29 +00:00
Code Issues Actions 4 Packages Projects Releases Wiki Activity

feat: Enhance Bedrock authentication with improved docs, validation, and tests

Browse Source 
- Add clearer authentication method comments in documentation examples
- Include security best practices for API key usage
- Add CLI validation hints for AWS Bedrock API key format
- Implement comprehensive error handling tests (timeout, rate limit, invalid key)
- Add connection error and retry scenario test cases
- Maintain consistency between English and Portuguese documentation

Addresses AI code review feedback on PR #3126

Co-Authored-By: Jo\u00E3o <joao@crewai.com>
This commit is contained in:
Devin AI
2025-07-09 12:24:56 +00:00
parent 3eea890409
commit f36d48d57e
4 changed files with 92 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
docs/en/concepts/llms.mdx
View File

@@ -328,12 +328,14 @@ In this section, you'll find detailed examples that help you select, configure,
Example usage in your CrewAI project:
```python Code
# Using IAM role authentication
# Method 1: IAM Role Authentication (uses AWS_ACCESS_KEY_ID + AWS_SECRET_ACCESS_KEY)
# Set environment variables: AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_DEFAULT_REGION
llm = LLM(
model="bedrock/anthropic.claude-3-sonnet-20240229-v1:0"
)
# Using API key authentication
# Method 2: API Key Authentication (uses AWS_BEARER_TOKEN_BEDROCK)
# Set environment variables: AWS_BEARER_TOKEN_BEDROCK, AWS_DEFAULT_REGION
llm = LLM(
model="bedrock/anthropic.claude-3-sonnet-20240229-v1:0"
)
@@ -344,6 +346,13 @@ In this section, you'll find detailed examples that help you select, configure,
- For API key authentication, you can generate a 30-day API key from the [Amazon Bedrock console](https://console.aws.amazon.com/bedrock/)
- For production applications, use IAM roles or temporary credentials instead of long-term API keys
**Security Best Practices:**
- API keys expire after 30 days and should be rotated regularly
- Use IAM roles for production environments for better security
- Store API keys securely and never commit them to version control
- Monitor API usage and set up alerts for unusual activity
- Consider using temporary credentials for enhanced security
[Amazon Bedrock](https://docs.aws.amazon.com/bedrock/latest/userguide/models-regions.html) is a managed service that provides access to multiple foundation models from top AI companies through a unified API, enabling secure and responsible AI application development.
| Model | Context Window | Best For |

13
docs/pt-BR/concepts/llms.mdx
View File

@@ -326,12 +326,14 @@ Nesta seção, você encontrará exemplos detalhados que ajudam a selecionar, co
Exemplo de uso em seu projeto CrewAI:
```python Code
# Usando autenticação por função IAM
# Método 1: Autenticação por Função IAM (usa AWS_ACCESS_KEY_ID + AWS_SECRET_ACCESS_KEY)
# Configure as variáveis de ambiente: AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_DEFAULT_REGION
llm = LLM(
model="bedrock/anthropic.claude-3-sonnet-20240229-v1:0"
)
# Usando autenticação por chave de API
# Método 2: Autenticação por Chave de API (usa AWS_BEARER_TOKEN_BEDROCK)
# Configure as variáveis de ambiente: AWS_BEARER_TOKEN_BEDROCK, AWS_DEFAULT_REGION
llm = LLM(
model="bedrock/anthropic.claude-3-sonnet-20240229-v1:0"
)
@@ -342,6 +344,13 @@ Nesta seção, você encontrará exemplos detalhados que ajudam a selecionar, co
- Para autenticação por chave de API, você pode gerar uma chave de 30 dias no [console do Amazon Bedrock](https://console.aws.amazon.com/bedrock/)
- Para aplicações de produção, use funções IAM ou credenciais temporárias em vez de chaves de API de longo prazo
**Melhores Práticas de Segurança:**
- Chaves de API expiram após 30 dias e devem ser rotacionadas regularmente
- Use funções IAM para ambientes de produção para melhor segurança
- Armazene chaves de API com segurança e nunca as confirme no controle de versão
- Monitore o uso da API e configure alertas para atividades incomuns
- Considere usar credenciais temporárias para segurança aprimorada
[Amazon Bedrock](https://docs.aws.amazon.com/bedrock/latest/userguide/models-regions.html) é um serviço gerenciado que fornece acesso a múltiplos modelos fundamentais dos principais provedores de IA através de uma API unificada, permitindo o desenvolvimento seguro e responsável de aplicações de IA.
| Modelo | Janela de Contexto | Melhor Para |

2
src/crewai/cli/constants.py
View File

@@ -63,7 +63,7 @@ ENV_VARS = {
"key_name": "AWS_REGION_NAME",
},
{
"prompt": "Enter your AWS Bedrock API Key (press Enter to skip)",
"prompt": "Enter your AWS Bedrock API Key (30-day key from AWS console, press Enter to skip)",
"key_name": "AWS_BEARER_TOKEN_BEDROCK",
},
],

69
tests/test_bedrock_authentication.py
View File

@@ -1,4 +1,5 @@
import os
import pytest
from unittest.mock import patch, MagicMock
from crewai import LLM
@@ -106,3 +107,71 @@ class TestBedrockAuthentication:
mock_completion.assert_called_once()
assert result == "Test response"
@patch.dict(os.environ, {
'AWS_BEARER_TOKEN_BEDROCK': 'test-api-key',
'AWS_DEFAULT_REGION': 'us-east-1'
})
@patch('litellm.completion')
def test_bedrock_timeout_handling(self, mock_completion):
"""Test Bedrock API timeout handling."""
mock_completion.side_effect = TimeoutError("Request timed out")
llm = LLM(model="bedrock/anthropic.claude-3-sonnet-20240229-v1:0")
with pytest.raises(TimeoutError, match="Request timed out"):
llm.call("test message")
@patch.dict(os.environ, {
'AWS_BEARER_TOKEN_BEDROCK': 'test-api-key',
'AWS_DEFAULT_REGION': 'us-east-1'
})
@patch('litellm.completion')
def test_bedrock_rate_limit_handling(self, mock_completion):
"""Test Bedrock API rate limit handling."""
mock_completion.side_effect = Exception("Rate limit exceeded")
llm = LLM(model="bedrock/anthropic.claude-3-sonnet-20240229-v1:0")
with pytest.raises(Exception, match="Rate limit exceeded"):
llm.call("test message")
@patch.dict(os.environ, {
'AWS_BEARER_TOKEN_BEDROCK': 'invalid-key',
'AWS_DEFAULT_REGION': 'us-east-1'
})
@patch('litellm.completion')
def test_bedrock_invalid_api_key(self, mock_completion):
"""Test Bedrock with invalid API key."""
mock_completion.side_effect = Exception("Invalid API key")
llm = LLM(model="bedrock/anthropic.claude-3-sonnet-20240229-v1:0")
with pytest.raises(Exception, match="Invalid API key"):
llm.call("test message")
@patch.dict(os.environ, {
'AWS_ACCESS_KEY_ID': 'test-key-id',
'AWS_SECRET_ACCESS_KEY': 'test-secret-key',
'AWS_DEFAULT_REGION': 'us-east-1'
})
@patch('litellm.completion')
def test_bedrock_connection_error(self, mock_completion):
"""Test Bedrock with connection error."""
mock_completion.side_effect = ConnectionError("Connection failed")
llm = LLM(model="bedrock/anthropic.claude-3-sonnet-20240229-v1:0")
with pytest.raises(ConnectionError, match="Connection failed"):
llm.call("test message")
@patch.dict(os.environ, {
'AWS_BEARER_TOKEN_BEDROCK': 'test-api-key',
'AWS_DEFAULT_REGION': 'us-east-1'
})
@patch('litellm.completion')
def test_bedrock_api_key_with_retry_scenario(self, mock_completion):
"""Test Bedrock API key authentication with retry scenario."""
mock_completion.side_effect = [
Exception("Temporary error"),
MagicMock(choices=[MagicMock(message=MagicMock(content="Success after retry"))])
]
llm = LLM(model="bedrock/anthropic.claude-3-sonnet-20240229-v1:0")
with pytest.raises(Exception, match="Temporary error"):
llm.call("test message")