AWS CloudWatch

Contents

• You get 10 CloudWatch metrics, 10 alarms, 1,000,000 API requests, and 1,000 Amazon SNS email notifications per customer per month for free.
• There is no limit on the number of custom metrics you can create.
• The maximum period you can specify is one day (86,400 seconds).
• You can assign up to 10 dimensions per metric.
• You can create up to 5000 alarms per AWS account.
• You can assign up to 5 actions per alarm.
• Metric data is kept for 2 weeks.
• The size of a PutMetricData request is limited to 8KB for HTTP GET requests and 40KB for HTTP POST requests.
• You can include a maximum of 20 MetricDatum items in one PutMetricData request. A MetricDatum can contain a single value or a StatisticSet representing many values.

• Up to 5GB of incoming data for free.
• Up to 5GB of data archiving for free.
• The maximum number of log groups per AWS account is 500.
• The maximum number of metric filters is 100 per log group.
• The maximum event size is 256KB.
• The maximum batch size is 1MB.
• The maximum rate of a PutLogEvents request is 5 requests per second per log stream. Since the maximum batch size of a PutLogEvents request is 32KB, this means that uploads to a single log stream are limited to a maximum rate of 160KB/s.
• The maximum rate of a GetLogEvents request is 10 requests per second per AWS account.

# pip install boto

[default]
aws_access_key_id=XXXXXXXXXXXXXXXXXXXXXXXX
aws_secret_access_key=XXXXXXXXXXXXXXXXXXXXXXXXXXXX

import boto.ec2.cloudwatch
import datetime

c = boto.ec2.cloudwatch.connect_to_region('ap-northeast-1')

# 최근 30분간의 데이터를 가져오고 싶다.
end = datetime.datetime.utcnow()
start = end - datetime.timedelta(minutes=30)

dimensions={'InstanceId':'i-xxxxxxxx'}
metric_name = 'CPUUtilization'
namespace = 'AWS/EC2'
statistics = ['Average']
unit = 'Percent'

datapoints = c.get_metric_statistics(300, start, end, metric_name,namespace , statistics, dimensions, unit);

for item in datapoints:
     print item['Average']

0.034
0.034
0.0
0.032
0.0

• metrics : 자원을 측정하기 위한 측정 기준 명. CPUUtilization(CPU 사용율), NetworkIn(네트워크 입력 트래픽)등이 metric이다.
• dimensions : metrics을 어느 기준(차원)으로 찾을 것인지를 설정하기 위해서 사용한다. 예를들어 EC2 매트릭은 ImageID, InstanceID, InstanceType등 다양한 기준으로 찾을 수 있다.

1. Zabbix 서버를 이용해서 모니터링한다.
2. Zabbix Agent는 CloudWatch API를 이용해서 주기적으로 각 매트릭을 수집한다.

• NumberOfMessageSent : 큐에 추가되는 메시지 갯수
• SentMessageSize : 큐에 추가된 메시지의 데이터 크기
• NumberOfMessgeReceves : 삭제한 메시지
• ApproximateNumberOfMessagesDelayed :
• ApproximateNumberOfMessagesVisible :
• approximateNumberOfMessageNotVisible :