OpenSearch Serverless: Search Without Cluster Management

OpenSearch Serverless: Search and Analytics Without Cluster Management

Managing Elasticsearch clusters is operational overhead that doesn’t add business value. OpenSearch Serverless eliminates nodes, shards, and capacity planning—you just query. In 2025, it’s become the default choice for new search workloads.

What OpenSearch Serverless Manages

Collection Types

# OpenSearch Serverless collection types

1. Search
   - Full-text search workloads
   - E-commerce product search
   - Document search
   - Optimized for query latency

2. Time Series
   - Log analytics
   - Metrics storage
   - IoT data
   - Optimized for time-based queries

3. Vector Search
   - Semantic search
   - RAG applications
   - Image similarity
   - k-NN queries

Creating a Serverless Collection

import boto3

client = boto3.client('opensearchserverless')

# Step 1: Create security policy
client.create_security_policy(
    name='products-encryption',
    type='encryption',
    policy='''
    {
        "Rules": [{"ResourceType": "collection", "Resource": ["collection/products"]}],
        "AWSOwnedKey": true
    }
    '''
)

# Step 2: Create network policy
client.create_security_policy(
    name='products-network',
    type='network',
    policy='''
    [{
        "Rules": [{"ResourceType": "collection", "Resource": ["collection/products"]}],
        "AllowFromPublic": true
    }]
    '''
)

# Step 3: Create collection
response = client.create_collection(
    name='products',
    type='SEARCH',
    description='Product catalog search'
)

print(f"Collection endpoint: {response['createCollectionDetail']['collectionEndpoint']}")

Vector Search for RAG

# Create vector index for RAG application
from opensearchpy import OpenSearch

client = OpenSearch(
    hosts=[{'host': 'your-collection.us-east-1.aoss.amazonaws.com', 'port': 443}],
    http_auth=awsauth,
    use_ssl=True
)

# Create index with k-NN
client.indices.create(
    index='knowledge-base',
    body={
        'settings': {
            'index.knn': True
        },
        'mappings': {
            'properties': {
                'embedding': {
                    'type': 'knn_vector',
                    'dimension': 1536,
                    'method': {'engine': 'faiss', 'name': 'hnsw'}
                },
                'text': {'type': 'text'},
                'metadata': {'type': 'object'}
            }
        }
    }
)