TestMemoryIntegration_* suite covering context assembly success, L1–L3 degradation fault injection, X-IBEX-Skip-Memory bypass, extraction enqueue semantics, and full Phase 1+2 regression.
Milestone 3.5.D.6 — Integration Tests
Status: Planned
Goal: Track D — Proxy Integration
Phase: 3.5 — Extraction & Context Assembly
Estimated effort: 2 days
Track: Track D — Proxy Integration
Depends on: Milestones 3.5.D.1–3.5.D.5
Why This Milestone Exists
Update the canonical middleware chain diagram to its final Phase 3.5 form and add TestMemoryIntegration_* suite in services/proxy, mirroring the TestSecurity_SEC* convention.
Test Coverage
- Context assembly success path with real memories present
- L1–L3 degradation paths (fault-injected via a test gRPC server that delays/errors on demand)
X-IBEX-Skip-Memorycorrectly bypasses the gRPC call (asserted via call-count mock)- Extraction enqueue fires exactly once per completed session, never on
activestatus - Full regression: all Phase 1 + Phase 2 security/session tests still pass with
IBEX_CONTEXT_ENABLED=true
Orientation (indicative)
Named paths, package layouts, libraries, schemas, env vars, and commands anywhere on this page are rough sketches for orientation — inspiration and a baseline, not a required change list.
During implementation, expect to:
- open the live tree and follow existing patterns before inventing new ones
- research current constraints (latency, tenancy, deploy shape, libraries) more deeply than this page can
- advance the design beyond the sketch where measurement or code reality says so
- land work in different filenames, merged packages, deferred docs, or new surfaces when the situation calls for it
Prefer outcomes over matching any particular file tree or command sequence.
Areas that may be involved (situational — not a checklist):
- Proxy service (HTTP, bootstrap, config)
- Context assembly service
- Workers / task runtime
Success signals
Outcome-oriented signals that the milestone is in good shape. Exact filenames, package layouts, and commands may differ from any sketches above.
- Fully unit-tested with fault injection for all 4 degradation levels
-
dispatchProviderCompletionwired, withIBEX_CONTEXT_ENABLED=falseverified to reproduce exact Phase 2 behavior - Both header-only and embedded-metadata response surfaces implemented
- Extraction enqueue uses Celery-native Redis list semantics
- Proxy-side benchmark confirms
<20msproxy overhead target with context assembly enabled - Repo guards / CI checks still pass
Prerequisites
- Milestones 3.5.D.1–3.5.D.5
Last updated on