调用Microsoft Graph API /communications/callRecords遇请求超时等问题求助

我搭建了Azure环境，尝试调用https://graph.microsoft.com/v1.0/communications/callRecords接口，一开始遇到未知错误，给应用开了全部API权限后，现在又碰到「请求超时错误」。附上我的代码（敏感信息已隐去）：

from msal import ConfidentialClientApplication
import requests
import json

# Replace these values with your actual application and authentication details
client_id = ' '
client_secret = ' '
tenant_id = ' '
authority = 'https://login.microsoftonline.com/' + tenant_id
scopes = ['https://graph.microsoft.com/.default']

# Set up MSAL ConfidentialClientApplication
app = ConfidentialClientApplication(
    client_id,
    authority=authority,
    client_credential=client_secret,
)

# Acquire a token
token_response = app.acquire_token_for_client(scopes=scopes)

# Set up the GraphServiceClient with the acquired token
access_token = token_response["access_token"]
headers = {
    "Authorization": f"Bearer {access_token}",
    "Content-Type": "application/json",
}

# Define the batch request payload
 
single_request = {
    "id": "1",
    "method": "GET",
    "url": "/communications/callRecords",
    # 考虑删除每个请求的不必要头部
}

# 拆分单个请求为20个子请求
batch_size = 20
requests_to_process = [single_request.copy() for _ in range(batch_size)]

batches = [requests_to_process[i:i + 1] for i in range(0, len(requests_to_process))]

for batch_number, batch in enumerate(batches, start=1):
    batch_payload = {"requests": batch}

    # 进行批量请求
    batch_response = requests.post(
        url="https://graph.microsoft.com/v1.0/$batch",
        headers=headers,
        json=batch_payload,
    )

    # 检查批量响应
    if batch_response.status_code != 200:
        print(f"Error in batch {batch_number}, status code: {batch_response.status_code}")
        print(batch_response.json())
    else:
        print(f"Batch {batch_number} response:")
        print(json.dumps(batch_response.json(), indent=4))

1. 批量请求逻辑完全错误 #

你当前的代码把20个完全相同的请求拆成了20个单请求批次，完全浪费了$batch批量请求的优势，反而会额外增加服务器连接开销，这是超时的核心原因之一。Graph API允许一个批次最多包含20个请求，应该把所有请求打包进同一个批次发送，而不是分20次单请求调用。

2. Call Records接口本身的性能问题 #

/communications/callRecords返回的数据量通常很大，直接无参数请求会导致响应缓慢甚至超时，建议：

• 添加$top参数限制单次返回的记录数，比如/communications/callRecords?$top=50
• 通过响应中的@odata.nextLink实现分页，分批获取全量数据

3. 增加超时时间设置 #

requests.post默认超时时间较短，可手动设置timeout参数（比如timeout=30，单位秒），避免因接口响应慢触发的超时错误。

4. 确认权限有效性 #

虽然你开了全部权限，但要确认应用是否拥有**CallRecords.Read.All权限**（这是调用该接口的必要权限），且该权限已获取管理员同意——仅添加权限未授予同意会导致权限无效。

5. 优化后的代码示例 #

from msal import ConfidentialClientApplication
import requests
import json

# 替换为你的实际信息
client_id = ' '
client_secret = ' '
tenant_id = ' '
authority = f'https://login.microsoftonline.com/{tenant_id}'
scopes = ['https://graph.microsoft.com/.default']

# 初始化MSAL客户端
app = ConfidentialClientApplication(
    client_id,
    authority=authority,
    client_credential=client_secret,
)

# 获取访问令牌（增加错误判断）
token_response = app.acquire_token_for_client(scopes=scopes)
if 'access_token' not in token_response:
    print(f"令牌获取失败: {token_response.get('error_description')}")
    exit()

access_token = token_response["access_token"]
headers = {
    "Authorization": f"Bearer {access_token}",
    "Content-Type": "application/json",
}

# 构建有效批量请求：一次发送20个分页请求（示例）
batch_requests = []
for i in range(20):
    # 按分页逻辑生成请求，根据实际需求调整参数
    request_url = f'/communications/callRecords?$top=50&$skip={i*50}'
    batch_requests.append({
        "id": str(i+1),
        "method": "GET",
        "url": request_url
    })

# 发送单个批量请求
try:
    batch_response = requests.post(
        url="https://graph.microsoft.com/v1.0/$batch",
        headers=headers,
        json={"requests": batch_requests},
        timeout=30  # 设置超时时间
    )
    batch_response.raise_for_status()  # 主动抛出HTTP错误
    print("批量请求响应:")
    print(json.dumps(batch_response.json(), indent=4))
except requests.exceptions.Timeout:
    print("请求超时，请检查网络或延长超时时间")
except requests.exceptions.RequestException as e:
    print(f"请求出错: {e}")
    if batch_response.status_code != 200:
        print(f"错误详情: {json.dumps(batch_response.json(), indent=4)}")

6. 其他注意事项 #

• 避免短时间内发送大量重复请求，Graph API有请求频次限制，过度请求会被限流
• 查看Azure AD应用的日志，获取更详细的错误排查信息
• 确认你的网络环境可以正常访问Graph API，排查防火墙、代理等限制

内容的提问来源于stack exchange，提问作者Muerta