majiayu000 Memory Pipeline E2E Test
Run the full embed → store → search pipeline:
Start port-forward (if not already running):
kubectl port-forward svc/qdrant -n vector 6333:6333 & sleep 2Generate embedding:
curl -s http://10.10.10.10:30001/v1/embeddings \ -H "Content-Type: application/json" \ -d '{"model":"default","input":"Kaizen memory pipeline test at '"$(date -Iseconds)"'"}'Verify: 4096 dimensions returned.
Store in Qdrant: Use the embedding from step 2 to upsert a point into the
episodiccollection with a random UUID and timestamp payload.Search: Generate a new embedding for a related query and search the
episodiccollection. Verify score > 0.5.Report:
Step Status Details Embedding ? dims, latency Store ? collection, point ID Search ? score, matches Cleanup port-forward:
kill %1 2>/dev/null