TypeError：リストのインデックスはstrでない整数でなければなりません

Question

キュー 'myDownlinQueue'のメトリック（トリガ）がApproximateNumberOfMessagesVisibleであるアラームを作成したい場合、アラームはAutoScalingGroupのオートスケーリングの動作を行います。しかし、私が書いているコードに基づいて、私はタイトルに記載されているエラーを取得します。たぶん私は間違っているが、以下はこのエラーの原因となるコードです。

# ==== AutoSCaling config ======================= 
autoscaling_group = { 
"name": "myAG", #descriptive name for your auto scaling group 
"min_size": 0 , #Minimum number of instances that should be running at all times 
"max_size": 0 #Maximum number of instances that should be running at all times 
} 


lc_name = 'myLG' #Descriptive name for your launch configuration 

#=================AMI to launch====================================================== 
as_ami = { 
"id": "ami-******c", #The AMI ID of the instance your Auto Scaling group will launch 
"VpcId" : "vpc-0c805575", 
"security_groups": "sg-xxxxxxxa", #The security group(s) your instances will belong to 
"instance_type": "t2.micro", #The size of instance that will be launched 
"instance_monitoring": True #Indicated whether the instances will be launched with detailed monitoring enabled. Needed to enable CloudWatch 
} 


conn_reg = boto.ec2.connect_to_region(region_name=awsRegion) 
zones = conn_reg.get_all_zones() 

zoneStrings = [] 
for zone in zones: 
    zoneStrings.append(zone.name) 
    print " Available zones : " + zone.name 

conn_vpc = boto.connect_vpc() 
subnetids = conn_vpc.get_all_subnets() 


conn_as = AutoScaleConnection(AWS_ACCESS_KEY,AWS_SECRET_KEY) 

lc = LaunchConfiguration(name = lc_name, 
         image_id = as_ami["id"], 
         instance_type = as_ami["instance_type"], 
         user_data = "user-data.bls", 
         associate_public_ip_address = True, 
         instance_monitoring=as_ami["instance_monitoring"]) 

conn_as.create_launch_configuration(lc) 

ag = AutoScalingGroup(group_name = autoscaling_group["name"], 
         availability_zones= zoneStrings, 
         vpc_zone_identifier = subnetList, 
         launch_config=lc, min_size = autoscaling_group["min_size"], 
         max_size = autoscaling_group["max_size"]) 
conn_as.create_auto_scaling_group(ag) 


#=================Create Scaling Policies=================================== 
# Policy for scaling the number of servers up and down 

scalingUpPolicy = ScalingPolicy(name = "myScaleUpPolicy", 
              adjustment_type ="ChangeInCapacity", 
              as_name=ag.name, 
              scaling_adjustment = numInstances , 
              cooldown=180) 

scalingDownPolicy = ScalingPolicy(name = "myScaleDownPolicy", 
              adjustment_type= "ExactCapacity", 
              as_name=ag.name, 
              scaling_adjustment= 0 , 
              cooldown=180) 

conn_as.create_scaling_policy(scalingUpPolicy) 
conn_as.create_scaling_policy(scalingDownPolicy) 


scalingUpPolicy = conn_as.get_all_policies(as_group="myAG", policy_names=["myScaleUpPolicy"])[0] 
scalingDownPolicy = conn_as.get_all_policies(as_group="myAG",policy_names=["myScaleDownPolicy"])[0] 


# =========== CloudWatch Connection ============= 
cw = connect_to_region(awsRegion)  

# =========== SNS Connection =================== 
sns = connect_to_region(awsRegion) 

cw = CloudWatchConnection(AWS_ACCESS_KEY,AWS_SECRET_KEY) 
sqs = SQSConnection(AWS_ACCESS_KEY,AWS_SECRET_KEY) 



# ==== Alarm ================= 
numberOfMessages = 1 
metric_object = cw.list_metrics(dimensions={"QueueName":"myDownlinkQueue"}, metric_name = "ApproximateNumberOfMessagesVisible",namespace = "AWS/SQS") 

alarm_name = "myAlarm" 

metric_object["ApproximateNumberOfMessagesVisible"].create_alarm(name =alarm_name, 
              comparison=">=", 
              threshold = numberOfMessages, 
              period = 60, 
              evaluation_periods = 1, 
              statistic = "Average", 
              alarm_actions=[scalingDownPolicy.policy_arn]) 

エラーが最後のメソッドである - アラームを作成し、正確scalingDownPolicy.policy_arn

Answer 1

どちらのOPは、ドキュメントを読み違えたり混乱させる。

boto2 cloudwatch.list_metric()によれば、項目はlistを返します。

list_metrics（next_tokenは=なし、寸法=なし、METRIC_NAME =なし、 名前空間=なし）

が有効なメトリックのリストを返すためにそこに記録されたデータ利用可能 が。

のでmetric_object["ApproximateNumberOfMessagesVisible"]はエラーTypeError : list indices must be integers not str

OTHを蹴るだろう、boto3辞書を返すcloudwatch.list_object。彼らは互換性がありません。ただし、ApproximateNumberOfMessagesVisibleを持つ辞書キーはありません。

{ 
    'Metrics': [ 
     { 
      'Namespace': 'string', 
      'MetricName': 'string', 
      'Dimensions': [ 
       { 
        'Name': 'string', 
        'Value': 'string' 
       }, 
      ] 
     }, 
    ], 
    'NextToken': 'string' 
}