Filter Out Unused AWS S3 Buckets based on threshold days, empty bucket and website configuration

import boto3 from datetime import datetime, timedelta, timezone from botocore.exceptions import ClientError creds = _get_creds(cred_label)['creds'] access_key = creds['username'] secret_key = creds['password'] # Initialize S3 client def get_s3_client(): return boto3.client('s3', aws_access_key_id=access_key, aws_secret_access_key=secret_key) def list_all_buckets(s3_client): try: buckets = s3_client.list_buckets() return buckets['Buckets'] except Exception as e: print(f"Failed to list S3 buckets: {e}") return [] def get_last_modified_object(s3_client, bucket_name): try: paginator = s3_client.get_paginator('list_objects_v2') page_iterator = paginator.paginate(Bucket=bucket_name) last_modified = None for page in page_iterator: if 'Contents' in page: for obj in page['Contents']: if last_modified is None or obj['LastModified'] > last_modified: last_modified = obj['LastModified'] return last_modified except Exception as e: print(f"Error accessing objects in bucket {bucket_name}: {e}") return None def is_bucket_empty(s3_client, bucket_name): response = s3_client.list_objects_v2(Bucket=bucket_name) return 'Contents' not in response def is_bucket_website(s3_client, bucket_name): try: s3_client.get_bucket_website(Bucket=bucket_name) return True except ClientError as e: if e.response['Error']['Code'] == 'NoSuchWebsiteConfiguration': return False raise # Rethrow the exception if it's not the expected "NoSuchWebsiteConfiguration" def filter_unused_buckets(s3_client, days_threshold): unused_buckets = [] current_time = datetime.now(timezone.utc) threshold_time = current_time - timedelta(days=days_threshold) buckets = list_all_buckets(s3_client) for bucket in buckets: bucket_name = bucket['Name'] if is_bucket_empty(s3_client, bucket_name) and not is_bucket_website(s3_client, bucket_name): last_modified = get_last_modified_object(s3_client, bucket_name) if last_modified is None or last_modified < threshold_time: unused_buckets.append(bucket_name) return unused_buckets # Main Execution s3_client = get_s3_client() #last_accessed_threshold_days = 90 # Define the threshold days unused_buckets = filter_unused_buckets(s3_client, last_accessed_threshold_days) #print(unused_buckets) # for debugging if unused_buckets: print(f"Unused buckets (no modifications for at least {last_accessed_threshold_days} days):") for bucket in unused_buckets: print(bucket) else: print("No unused buckets found based on the specified threshold.")