Read AWS CloudWatch metrics from boto3

For several reasons, I need to read from Python boto3 the number of Requests on my Target Group.

I need to have the 186 number in my boto3 Lambda to run others checks and actions.

When I started this I thought it was an easy one but it wasn’t like that. So I’m writing this very short guide to save you and to my future self :)

GetMetricData or GetMetricStatistics

If you go to the CloudWatch page for boto3 you found 2 options

• get_metric_data()
• get_metric_statistics()

I tested both and to be sure to do the right choice I ask to AWS Premium support what is the right choice. They point me to this documentation page where there is a good explanation and also some example with AWS CLI, this really helps me to chose the metric_data but it wasn’t enough because I need to have it in boto3.

Solution

This is my final solution a function that returns the final number 186 in my initial example.

I used a filter based on days because it was enough for my purpose but if you are a good python developer (I’m not) you can customize using also hours.

I’m taking back all the results from the end of the current day (because when you add 1 day it starts from midnight) and a number of days in the past, in the example it was 2.

import boto3
import os
from datetime import date, timedelta, datetime
cloudwatch_client = boto3.client('cloudwatch')
def handler(event, context):
  mytargetgrouparn='arn:aws:elasticloadbalancing:eu-west-1:012345678:targetgroup/mytgname/d3cf755348f4b85d'
  myapplicationlbarn='arn:aws:elasticloadbalancing:eu-west-1:012345678:loadbalancer/app/myappliclbname/11229988hfhunls'    
 print(access_last_days_data_metric(mytargetgrouparn,myapplicationlbarn,2))
def access_last_days_data_metric(targetGroupARN,loadBalancerARN,days):
    tgarray=targetGroupARN.split(':')
    tgstring=tgarray[-1]
    
    lbarray=loadBalancerARN.split(':')
    lbstring=lbarray[-1]
    lbarray2=lbstring.split('/')
    lbstring2=lbarray2[1]+'/'+lbarray2[2]+'/'+lbarray2[3]
    yesterday=date.today() - timedelta(days=days)
    tomorrow=date.today() + timedelta(days=1)
    
    
    response = cloudwatch_client.get_metric_data(
        MetricDataQueries=[
            {
                'Id': 'myrequest',
                'MetricStat': {
                    'Metric': {
                        'Namespace': 'AWS/ApplicationELB',
                        'MetricName': 'RequestCount',
                        'Dimensions': [
                            {
                                'Name': 'TargetGroup',
                                'Value': tgstring
                            },
                            {                        
                                'Name': 'LoadBalancer',
                                'Value': lbstring2
                            },
                        ]
                    },
                    'Period': 300,
                    'Stat': 'Sum',
                    'Unit': 'Count'
                }
            },
        ],
        StartTime=datetime(yesterday.year, yesterday.month, yesterday.day), 
        EndTime=datetime(tomorrow.year, tomorrow.month, tomorrow.day),    
    )
return sum(response['MetricDataResults'][0]['Values'])

If you want to know more about Python date and time objects here a good article

Feedback

If you like this article, and you want to motivate me to continue to write, please:

• leave a comment on the article
• clap here in medium
• connect with me on LinkedIn
• subscribe to my youtube channel AWS Tutorials — Giuseppe Borgese