Update endpoint

documented

Update an Inference Endpoint (e.g., min/max replicas, repo, hardware).

PATCH /endpoints/{endpointName}Inference Endpointsno implementation yet0 runs

Runs (30 d)

Success rate

—

p50 latency

—

p95 latency

—

Distinct callers

Runs per day

No run in the last 30 days.

Every run is logged with a redacted input and an anonymised output shape. Stages promote automatically: developed after one good run, tested when benchmarks pass, reliable at ≥97% over 30 days across several callers.