Fetch Application Inventory using Systems Manager

Background

We need a way to report on “software versions” deployed to Cloud.

Objective

Get application inventory from each EC2 instance and store it in centralized account S3 bucket. To query the information from s3 bucket, we are integrating Glue, Athena (from another account).

Proposed Architecture

Implementation Procedure

Enable system manager inventory in (PROD/NONPROD) each account (one-time activity)

Create S3 bucket in centralised account to receive logs from system manager (one-time activity)

Create resource data sync in each account (PROD/NONPROD) to send the log to centralised bucket (one-time activity)

Create Lambda function in centralised account to assign object level permission to each object if the analytical services are running in different account, so object level permission needs to be updated (work around given by AWS)

Configure Glue and Athena in required account to query the S3 bucket (s3://ssm-application-inventory-logs).

Enable System Manager inventory in each account

Its a one time activity in each account , we have to enable this feature using AWS console.

[https://docs.aws.amazon.com/systems-manager/latest/userguide/sysman-inventory-configuring.html]

Glue Details

Create Glue crawler, Glue database and Table in required account to match the ssm inventory output.

Athena Query

This service is enabled in a different account.

Run Crawler For Latest Report

To get latest report we have to run Glue crawler manually. Post successful execution, follow the Athena examples.

EXAMPLE 1:

To fetch whole inventory details from all the accounts, use the below query.

SELECT * from aws_application

EXAMPLE 2:

To fetch application wise instance details use the below query — replace ‘java%’ with your application name.

SELECT java.name, java.version, a.resourceid, accountid 
FROM “ssm-inventory-crawler-db”.”aws_application” a, (select distinct (resourceid), version, name from “ssm-inventory-crawler-db”.”aws_application” where name like ‘java%’) java
where a.resourceid = java.resourceid
order by a.resourceid

EXAMPLE 3:

To fetch custom inventory library like “python” using Athena,

SELECT * FROM “ssm-inventory-crawler-db”.”custom_python2libraries”;

EXAMPLE 4:

To fetch custom inventory library like “npm” using Athena,

SELECT * FROM “ssm-inventory-crawler-db”.”custom_nodejsnpm”;

Run Query in an automated way using Lambda

You can use below python script to query Athena every month day 1 using CloudWatch event rule and Lambda.

#!/usr/bin/python3
import boto3
import datetime
import os
import json
import logging
# boto3.setup_default_session(profile_name='saml',region_name='ap-southeast-1')
logger = logging.getLogger()
logger.setLevel(logging.DEBUG)
def poll_status(_id, message):
    athena= boto3.client('athena',region_name='ap-southeast-1')
    result = athena.get_query_execution( QueryExecutionId = _id )
    state  = result['QueryExecution']['Status']['State']

    while state == "QUEUED":
        result = athena.get_query_execution( QueryExecutionId = _id )
        state  = result['QueryExecution']['Status']['State']
        while state == "RUNNING":
            result = athena.get_query_execution( QueryExecutionId = _id )
            state  = result['QueryExecution']['Status']['State']
            if state == "FAILED":
                logger.info("FAILED-LOG: {} ".format(result))
        # return result
    else:
        logger.info("SUCCESS-LOG : {} ".format(result))
        sns_arn = os.environ["SNS_ARN"] #"arn:aws:sns:ap-southeast-1:928167378288:application-inventory-logs" 
        notification= _sns_notification(sns_arn,message)
        return result
def run_query(query, database, s3_output):
    client = boto3.client('athena',region_name='ap-southeast-1')
    response = client.start_query_execution(
        QueryString=query,
        QueryExecutionContext={
            'Database': database
            },
        ResultConfiguration={
            'OutputLocation': s3_output,
            }
        )
    print('Execution ID: ' + response['QueryExecutionId'])
    QueryExecutionId = response['QueryExecutionId']
    return QueryExecutionId

        # local_filename = QueryExecutionId + '.csv'
def _sns_notification(snsarn,message):
    client = boto3.client('sns')
    response = client.publish(TargetArn=snsarn,Message=json.dumps({'default': json.dumps(message)}),MessageStructure='json')
    logger.warning("SNS-NOTIFICATION : {} ".format(response))
    return response
def lambda_handler(event, context):
    #Athena configuration
    s3_output = os.environ['ATHENA_OUTPUT_BUCKET'] #'s3://athena-query-result-logs/result/' 
    database =   os.environ['GLUE_DB'] #'ssminventorycrawlerdb'
    table = os.environ["GLUE_TABLE"] #'aws_application' #
    app_search = os.environ["APPLICATION_LIST"] # "java% python% node%" #
    app_list= app_search.split(' ')
    sns_arn = "arn:aws:sns:ap-southeast-1:928167378288:application-inventory-logs" #os.environ["SNS-ARN"]
    #Athena database and table definition

    #Query definitions & Execute all queries
    python_library = "SELECT * FROM "'"%s"'"."'"%s"'" ;" % (database, "custom_python2libraries")
    npm_librarary = "SELECT * FROM "'"%s"'"."'"%s"'" ;" % (database, "custom_nodejsnpm")
    all_app = "SELECT * FROM "'"%s"'"."'"%s"'" ;" % (database, table)
    queries = [ python_library, npm_librarary, all_app ]
    for qa in queries:
        print("Executing query: %s" % (qa))
        resp = run_query(qa, database, s3_output)
        logger.info("QUERY-ID: {} ".format(resp))
        s3_key = resp + '.csv'
        message= {"Application-Inventory": s3_key,"Bucket-Details": s3_output,"App-name":qa}
        logger.info("MESSAGE-DETAILS : {} ".format(message))
        result = poll_status(resp, message)
    for app in app_list:
        specific_app = ("""SELECT java.name, java.version, a.resourceid, accountid 
                    FROM "%s"."%s" a, (select distinct (resourceid), version, name from "%s"."%s" where name like '%s') java
                    where a.resourceid = java.resourceid
                    order by a.resourceid """) % (database, table,database, table,app)
        # queries = [ python_library, npm_librarary, all_app, specific_app ]
        # for q in queries:
        print("Executing query: %s" % (specific_app))
        res = run_query(specific_app, database, s3_output)
        logger.info("QUERY-ID: {} ".format(res))
        s3_key = res + '.csv'
        messages= {"Application-Inventory": s3_key,"Bucket-Details": s3_output,"App-name":specific_app}
        logger.info("MESSAGE-DETAILS : {} ".format(messages))
        results = poll_status(res, messages)

if __name__ == '__main__':
    lambda_handler("event", "context")