python - 无法筛选 aws lambda 函数的 boto3 日志流

Question

我无法使用 boto3 API 和 'endtime' 过滤器过滤出 lambda 函数的 logStreams或'logStreamNamePrefix'过滤器。

这有效:

client = boto3.client('logs',
                      aws_access_key_id=aws_account['access_key'],
                      aws_secret_access_key=aws_account['secret_key'],
                      region_name=region)
logGroupName = '/aws/lambda/' + function_name
response = client.filter_log_events(
    logGroupName=logGroupName,
    startTime=int((datetime(2020, 1, 15)- datetime(1970, 1, 1)).total_seconds()),   # epoch_time        
)

但这些没有:

1.

response = client.filter_log_events(
    logGroupName=logGroupName,
    startTime=int((datetime(2020, 1, 15)- datetime(1970, 1, 1)).total_seconds()),   # epoch_time    
    endTime=int((datetime(2020, 1, 31)- datetime(1970, 1, 1)).total_seconds())       
)

ERROR: Empty list is returned, although log data is there for the specified time range

2.

response = client.filter_log_events(
    logGroupName=logGroupName,
    startTime=int((datetime(2020, 1, 15)- datetime(1970, 1, 1)).total_seconds()),   # epoch_time    
    logStreamNamePrefix='^2020/01'  # logs starting with this prefix        
)

ERROR:

botocore.exceptions.ParamValidationError: Parameter validation failed:
Unknown parameter in input: "logStreamNamePrefix", must be one of: logGroupName, logStreamNames, startTime, endTime, filterPattern, nextToken, limit, interleaved

Answer 1

作为临时修复，我正在轮询所有响应，然后明确过滤它。

response = client.filter_log_events(
    logGroupName=logGroupName,
    startTime=int((datetime(2020, 1, 15)- datetime(1970, 1, 1)).total_seconds()),   # epoch_time
)

filtered_response = list(filter(lambda x: x['logStreamName'].startswith('2020/01'), response['events']))