testOps

testOps is a tool to deploy, time and analyse FaaS functions currently AWS and GCP are supported

dependencies

• python3.9+
• virtualenv installed
• lockfile for usage with venv is supplied

installation

• create virtualenviroment
• navigate to venv\Scripts and run activate (depending on shell and os)
• then use pipenv sync to install dependencies (in testOps root folder)

usage

python testops.py [-h] [-d] [-i] [-a] [-keep {all,none,pareto}] filename

• -h displays help
• -d or --deploy activates the deployer functionality
• -i or --invoke activates the invoker and timing functionality
• -a or --analyse activates the analyser functionality
• -keep pick all, none or pareto to declare which functions should be kept after the run is finished, the pareto option requires -a (the analyser) to be active. if this option is ignores -keep all is used
• filename the filename of the input json

requirements

• testops requires a deployed pyStorage function on AWS
• testops requires a credentials.json which holds the credentials for AWS / GCP

example credentials json

{  
  "amazon":  
  {  
    "aws_access_key_id": "xyz",  
    "aws_secret_access_key": "xyz",  
    "aws_session_token": "xyz"  
  },  
  "google":  
  {  
    "client_email": "[email protected]",  
    "private_key":"-----BEGIN PRIVATE KEY-----\xyz\n-----END PRIVATE KEY-----\n",  
    "project_id": "xyz"  
  }  
}

example input json

{  
 "function_name": "function_name",  
 "aws_code": "s3://xyz.zip",  
 "gcp_code": "s3://xyz.zip",  
 "no_op_code": "s3://xyz.zip",  
 "aws_handler": "lambda_function.lambda_handler",  
 "gcp_handler":  "entry_handler",  
 "no_op_handler_aws": "lambda_function.lambda_handler",  
 "no_op_handler_gcp":  "entry_handler",  
 "gcp_project_id": "xyz",    
 "aws_deployment_role" : "arn:/LabRole",  
 "pyStorage_arn": "arn::function:pyStorage",  
 "aws_runtime": "python3.9",  
 "gcp_runtime": "python39",  
 "file_type": "zip",  
 "repetitions_of_experiment": 2,  
    "repetitions_per_function": 6,  
    "concurrency": 3,  
 "payload": [{"input": 25}],  
 "AWS_regions": {  
  "us-west-2": {  
   "memory_configurations":[  
    128  
  ]},  
  "us-east-1":{  
   "memory_configurations" :[  
    226  
  ]}
 },
 "GCP_regions": {  
  "us-west2": {  
   "memory_configurations": [  
    128  
   ]}  
 }
}

aws layer creator for python

small bonus, in folder aws layer creator you can find a docker script to create an AWS layer out of a requirement.txt refer to readme in the folder