Skip to content
InnoQualis InnoQualis Docs

Runbook — QdrantDown

Alert

Section titled “Alert”

QdrantDown

Severity

Section titled “Severity”

critical

Audience

Section titled “Audience”

ops

Owner

Section titled “Owner”

Phase 30.3 / KAN-293

What this alert will do

Section titled “What this alert will do”

Qdrant runs embedded in-process inside the backend (see CLAUDE.md — “Qdrant (embedded, in-process)”). When its index files are corrupted or unreachable, RAG / vector search returns errors. The alert fires when the backend reports vector-store unhealthy via /api/status for 2+ minutes.

Fallback procedure (until 30.3 lands)

Section titled “Fallback procedure (until 30.3 lands)”
Terminal window
# Backend's view of Qdrant health:
docker exec innoqualis-backend curl -fsS http://localhost:8000/api/status | jq '.checks.vector_store'


# Inspect the embedded Qdrant data directory:
docker exec innoqualis-backend ls -la /app/data/qdrant


# Restart the backend (Qdrant reinitialises on backend startup):
docker compose -f docker/compose.yml restart backend

If reindexing is needed: docker exec innoqualis-backend python -m app.ai.reindex_all (long-running; AI features stay degraded until complete).

Section titled “Related”
  • Phase 30.3 spec
  • backend/app/ai/vector_store.py — embedded Qdrant client

Last reviewed

Section titled “Last reviewed”

2026-05-29 — ops (stub)