Parallel Execution¶
The strategy used here will be to execute Prowler once per service. You can modify this approach as per your requirements.
This can help for really large accounts, but please be aware of AWS API rate limits:
- Service-Specific Limits: Each AWS service has its own rate limits. For instance, Amazon EC2 might have different rate limits for launching instances versus making API calls to describe instances.
- API Rate Limits: Most of the rate limits in AWS are applied at the API level. Each API call to an AWS service counts towards the rate limit for that service.
- Throttling Responses: When you exceed the rate limit for a service, AWS responds with a throttling error. In AWS SDKs, these are typically represented as
ThrottlingException
orRateLimitExceeded
errors.
For information on Prowler's retrier configuration please refer to this page.
Note
You might need to increase the --aws-retries-max-attempts
parameter from the default value of 3. The retrier follows an exponential backoff strategy.
Linux¶
Generate a list of services that Prowler supports, and populate this info into a file:
Make any modifications for services you would like to skip scanning by modifying this file.
Then create a new PowerShell script file parallel-prowler.sh
and add the following contents. Update the $profile
variable to the AWS CLI profile you want to run Prowler with.
#!/bin/bash
# Change these variables as needed
profile="your_profile"
account_id=$(aws sts get-caller-identity --profile "${profile}" --query 'Account' --output text)
echo "Executing in account: ${account_id}"
# Maximum number of concurrent processes
MAX_PROCESSES=5
# Loop through the services
while read service; do
echo "$(date '+%Y-%m-%d %H:%M:%S'): Starting job for service: ${service}"
# Run the command in the background
(prowler -p "$profile" -s "$service" -F "${account_id}-${service}" --only-logs; echo "$(date '+%Y-%m-%d %H:%M:%S') - ${service} has completed") &
# Check if we have reached the maximum number of processes
while [ $(jobs -r | wc -l) -ge ${MAX_PROCESSES} ]; do
# Wait for a second before checking again
sleep 1
done
done < ./services
# Wait for all background processes to finish
wait
echo "All jobs completed"
Output will be stored in the output/
folder that is in the same directory from which you executed the script.
Windows¶
Generate a list of services that Prowler supports, and populate this info into a file:
prowler aws --list-services | ForEach-Object {
# Capture lines that are likely service names
if ($_ -match '^\- \w+$') {
$_.Trim().Substring(2)
}
} | Where-Object {
# Filter out empty or null lines
$_ -ne $null -and $_ -ne ''
} | Set-Content -Path "services"
Make any modifications for services you would like to skip scanning by modifying this file.
Then create a new PowerShell script file parallel-prowler.ps1
and add the following contents. Update the $profile
variable to the AWS CLI profile you want to run prowler with.
Change any parameters you would like when calling prowler in the Start-Job -ScriptBlock
section. Note that you need to keep the --only-logs
parameter, else some encoding issue occurs when trying to render the progress-bar and prowler won't successfully execute.
$profile = "your_profile"
$account_id = Invoke-Expression -Command "aws sts get-caller-identity --profile $profile --query 'Account' --output text"
Write-Host "Executing Prowler in $account_id"
# Maximum number of concurrent jobs
$MAX_PROCESSES = 5
# Read services from a file
$services = Get-Content -Path "services"
# Array to keep track of started jobs
$jobs = @()
foreach ($service in $services) {
# Start the command as a job
$job = Start-Job -ScriptBlock {
prowler -p ${using:profile} -s ${using:service} -F "${using:account_id}-${using:service}" --only-logs
$endTimestamp = Get-Date -Format "yyyy-MM-dd HH:mm:ss"
Write-Output "${endTimestamp} - $using:service has completed"
}
$jobs += $job
Write-Host "$(Get-Date -Format 'yyyy-MM-dd HH:mm:ss') - Starting job for service: $service"
# Check if we have reached the maximum number of jobs
while (($jobs | Where-Object { $_.State -eq 'Running' }).Count -ge $MAX_PROCESSES) {
Start-Sleep -Seconds 1
# Check for any completed jobs and receive their output
$completedJobs = $jobs | Where-Object { $_.State -eq 'Completed' }
foreach ($completedJob in $completedJobs) {
Receive-Job -Job $completedJob -Keep | ForEach-Object { Write-Host $_ }
$jobs = $jobs | Where-Object { $_.Id -ne $completedJob.Id }
Remove-Job -Job $completedJob
}
}
}
# Check for any remaining completed jobs
$remainingCompletedJobs = $jobs | Where-Object { $_.State -eq 'Completed' }
foreach ($remainingJob in $remainingCompletedJobs) {
Receive-Job -Job $remainingJob -Keep | ForEach-Object { Write-Host $_ }
Remove-Job -Job $remainingJob
}
Write-Host "$(Get-Date -Format 'yyyy-MM-dd HH:mm:ss') - All jobs completed"
Output will be stored in C:\Users\YOUR-USER\Documents\output\
Combining the output files¶
Guidance is provided for the CSV file format. From the ouput directory, execute either the following Bash or PowerShell script. The script will collect the output from the CSV files, only include the header from the first file, and then output the result as CombinedCSV.csv in the current working directory.
There is no logic implemented in terms of which CSV files it will combine. If you have additional CSV files from other actions, such as running a quick inventory, you will need to move that out of the current (or any nested) directory, or move the output you want to combine into its own folder and run the script from there.
#!/bin/bash
# Initialize a variable to indicate the first file
firstFile=true
# Find all CSV files and loop through them
find . -name "*.csv" -print0 | while IFS= read -r -d '' file; do
if [ "$firstFile" = true ]; then
# For the first file, keep the header
cat "$file" > CombinedCSV.csv
firstFile=false
else
# For subsequent files, skip the header
tail -n +2 "$file" >> CombinedCSV.csv
fi
done
# Get all CSV files from current directory and its subdirectories
$csvFiles = Get-ChildItem -Recurse -Filter "*.csv"
# Initialize a variable to track if it's the first file
$firstFile = $true
# Loop through each CSV file
foreach ($file in $csvFiles) {
if ($firstFile) {
# For the first file, keep the header and change the flag
$combinedCsv = Import-Csv -Path $file.FullName
$firstFile = $false
} else {
# For subsequent files, skip the header
$tempCsv = Import-Csv -Path $file.FullName
$combinedCsv += $tempCsv | Select-Object * -Skip 1
}
}
# Export the combined data to a new CSV file
$combinedCsv | Export-Csv -Path "CombinedCSV.csv" -NoTypeInformation
TODO: Additional Improvements¶
Some services need to instantiate another service to perform a check. For instance, cloudwatch
will instantiate Prowler's iam
service to perform the cloudwatch_cross_account_sharing_disabled
check. When the iam
service is instantiated, it will perform the __init__
function, and pull all the information required for that service. This provides an opportunity for an improvement in the above script to group related services together so that the iam
services (or any other cross-service references) isn't repeatedily instantiated by grouping dependant services together. A complete mapping between these services still needs to be further investigated, but these are the cross-references that have been noted:
- inspector2 needs lambda and ec2
- cloudwatch needs iam
- dlm needs ec2