Azure OpenAI Embeddings
Embedding models served by Azure OpenAI.
Problem it solves: Semantic matching needs a numeric representation.
Where it sits: Before the vector database on query and ingest paths.
When to use
- RAG, semantic search, dedup.
When not to use
- Exact-match lookups.
Failure modes
- Model upgrade invalidates stored vectors (re-embed).
Metrics to watch
- Embeddings/s
- Latency
Alternatives
- Keyword search
Configuration in Labs (illustrative defaults)
- Instances (count) = 2 — Identical copies behind this logical component. Capacity scales linearly in this model.
- Capacity per instance (req/s) = 500 — Maximum requests one instance completes per second before it saturates.
- Base service time (ms) = 20 — Processing time per request at zero load. Queueing adds to this as utilisation rises.
Vendor examples (pattern only): Azure managed service. Simulated as a generic “Embedding model”.