Qdrant Component

Vector similarity search engine for AI/ML applications (RAG, semantic search, recommendations).

Architecture

                Qdrant Server - Vector database

                Collections - Vector storage units (via sub-component)

                Init Job - Auto-creates collections on startup

Quick Reference

REQUIRED = Must be defined by user

Attribute	Example	Default	Effect
`namespace` REQ	`qdrant`	-	K8s namespace
`global_version`	`1.16.2`	1.16.2	Qdrant version
`service_port`	`6333`	6333	REST API port
`cpu_request / cpu_limit`	`500m / 2000m`	-	CPU resources
`mem_request / mem_limit`	`1Gi / 4Gi`	-	Memory resources

Link Variables

Variable	Link Type	Purpose
`__prometheus`	prometheus-qdrant	Metrics collection
`__collection`	(sub-component)	Collection definitions

Sub-Components: collection

                collection_name - Name in Qdrant (default: sub-component name)

                vector_size - Vector dimensions (must match embedding model)

                distance - Cosine (default), Euclid, Dot

                on_disk - Store vectors on disk (bool)

                replication_factor - Replicas (default: 1)

                shard_number - Shards (default: 1)

Common Vector Sizes

                BAAI/bge-small-en-v1.5: 384

                BAAI/bge-base-en-v1.5: 768

                BAAI/bge-large-en-v1.5: 1024

                Alibaba-NLP/gte-Qwen2-1.5B-instruct: 1536

                OpenAI text-embedding-3-small: 1536

                OpenAI text-embedding-3-large: 3072

Generated Files

File	Condition	Contains
helm/helm-values.yaml	Always	Helm chart config
init-collections.yaml	__collection exists	K8s Job to create collections

Ports

Port	Purpose	Protocol
6333	REST API	HTTP
6334	gRPC API	gRPC
6335	gRPC companion	gRPC