pylambda_cputrace

Lambda Layers to measure cpu in Python3 applications.

How to use

Create Lambda layers

make package S3_BACKET='backet_name' S3_PREFIX='prefix_name'
make deploy STACK_NAME='stack_name'

Note: make package uses the dokcer image of pylambda-packer for build. It is because some Python module requires build on AmazonLinux.

Use with the Lambda function

Add AWSLambda-Python3-TraceUtil to Layers of your Lambda function. Import the package and add the decorator in the lambda function.

import json
from traceutil import cpu_trace
import time 

@cpu_trace
def lambda_handler(event, context):
    time.sleep(1)

    # TODO implement
    return {
        'statusCode': 200,
        'body': json.dumps('Hello from Lambda!')
    }

The CloudWatch Log shows the CPU utilization when running Lambda.

{'user': 0.0510204081633397, 'nice': 0.0, 'system': 0.010204081632644737, 'idle': 0.9285714285713672, 'iowait': 0.0, 'irq': 0.0, 'softirq': 0.005102040816322369, 'steal': 0.005102040816325994, 'guest': 0.0, 'guest_nice': 0.0}