🏥 Quick Return to Emergency Room
You are in a specialist desk.
For full triage and doctors on duty, return here:
- WFGY Global Fix Map — main Emergency Room, 300+ structured fixes
- WFGY Problem Map 1.0 — 16 reproducible failure modes
Think of this page as a sub-room.
If you want full consultation and prescriptions, go back to the Emergency Room lobby.
A compact hub to ship safely and keep RAG/LLM systems stable after release.
Use this folder to pick the right guardrail, verify with measurable targets, and recover fast when things wobble. No infra change required.
- Visual recovery map → RAG Architecture & Recovery
- Retrieval knobs end-to-end → Retrieval Playbook
- Traceability and snippet schema → Retrieval Traceability · Data Contracts
- Boot order and deploy traps → Bootstrap Ordering · Deployment Deadlock · Pre-Deploy Collapse
- Live ops tools → Live Monitoring for RAG · Debug Playbook
- First calls after deploy crash or return stale content.
- ΔS and citations look fine yesterday but flip today.
- Rate limits cascade, queues spike, latency climbs.
- Canary looks good then full rollout breaks retrieval.
- Index swap succeeds but answers cite old snippets.
- Retries cause duplicate side effects or charges.
- Feature flags bleed traffic into unfinished paths.
- Maintenance windows corrupt embeddings or anchors.
- ΔS(question, retrieved) ≤ 0.45 across three paraphrases.
- Coverage ≥ 0.70 on the expected new section.
- λ remains convergent on 2 seeds during rollout.
- Idempotency ≥ 99.9% on retry storms.
- Zero silent index mismatches (hash + counts match).
- P95 latency stays in budget with backpressure active.
| Scenario | Fix Page |
|---|---|
| Rollout readiness | rollout_readiness_gate.md |
| Canary strategy | staged_rollout_canary.md |
| Blue/green cutover | blue_green_switchovers.md |
| Version pin & freeze | version_pinning_and_model_lock.md |
| Vector index swap | vector_index_build_and_swap.md |
| Cache warmup | cache_warmup_invalidation.md |
| Rate limits | rate_limit_backpressure.md |
| Feature flags | feature_flags_safe_launch.md |
| Idempotency | idempotency_dedupe.md |
| Retry logic | retry_backoff.md |
| Rollback plan | rollback_and_fast_recovery.md |
| Postmortems | postmortem_and_regression_tests.md |
| Change freeze | release_calendar_and_change_freeze.md |
| Incident comms | incident_comms_and_statuspage.md |
| Shadow traffic | shadow_traffic_mirroring.md |
| Maintenance window | read_only_mode_and_maintenance_window.md |
| DB migrations | db_migration_guardrails.md |
- Freeze the world → Pin model IDs, prompt revs, index hashes.
- Warm up safely → Build index off-path, preload caches with canary.
- Shadow then canary → Mirror prod queries, step rollout 5% → 25% → 100%.
- Guard the edge → Enable backpressure, retries with jitter, idempotency keys.
- Know your exit → Keep rollback switch and comms draft ready.
| What you see | Open this |
|---|---|
| Deploy points to old snippets | vector_index_build_and_swap.md · cache_warmup_invalidation.md |
| Canary fine, full rollout breaks | staged_rollout_canary.md · feature_flags_safe_launch.md |
| Wrong model after failover | version_pinning_and_model_lock.md |
| Retries duplicate charges | idempotency_dedupe.md · retry_backoff.md |
| RL storms, timeouts | rate_limit_backpressure.md |
| Need rollback now | rollback_and_fast_recovery.md · blue_green_switchovers.md |
| Maintenance corrupts anchors | read_only_mode_and_maintenance_window.md · db_migration_guardrails.md |
| Unsure if safe to ship | rollout_readiness_gate.md |
Q: What does ΔS mean here?
A: ΔS is a stability score. It measures how much the retrieved content drifts from the expected anchor when you change the query slightly. Lower is better (≤ 0.45 is safe).
Q: What is λ convergence?
A: λ tracks whether retrieval order flips unpredictably. If λ is stable across seeds, your rollout is consistent.
Q: Why do I need idempotency keys?
A: Without them, retries can double-charge a user or run the same side-effect twice. Keys make every request “safe to retry.”
Q: How do I know if my index swap worked?
A: Check doc counts and hashes before cutover. If they mismatch, you’re pointing at an incomplete index.
Q: Canary looked fine but production broke — why?
A: Canary often hides tail-latency, cache misses, or load-based rate limits. Always test at increasing % of live traffic.
Q: Why do you mention rollback comms?
A: Technical rollback is only half. Users and stakeholders need fast updates, so pre-draft Statuspage or Slack messages are essential.
| Tool | Link | 3-Step Setup |
|---|---|---|
| WFGY 1.0 PDF | Engine Paper | 1️⃣ Download · 2️⃣ Upload to your LLM · 3️⃣ Ask “Answer using WFGY + <your question>” |
| TXT OS (plain-text OS) | TXTOS.txt | 1️⃣ Download · 2️⃣ Paste into any LLM chat · 3️⃣ Type “hello world” — OS boots instantly |
| Layer | Page | What it’s for |
|---|---|---|
| ⭐ Proof | WFGY Recognition Map | External citations, integrations, and ecosystem proof |
| ⚙️ Engine | WFGY 1.0 | Original PDF tension engine and early logic sketch (legacy reference) |
| ⚙️ Engine | WFGY 2.0 | Production tension kernel for RAG and agent systems |
| ⚙️ Engine | WFGY 3.0 | TXT based Singularity tension engine (131 S class set) |
| 🗺️ Map | Problem Map 1.0 | Flagship 16 problem RAG failure taxonomy and fix map |
| 🗺️ Map | Problem Map 2.0 | Global Debug Card for RAG and agent pipeline diagnosis |
| 🗺️ Map | Problem Map 3.0 | Global AI troubleshooting atlas and failure pattern map |
| 🧰 App | TXT OS | .txt semantic OS with fast bootstrap |
| 🧰 App | Blah Blah Blah | Abstract and paradox Q&A built on TXT OS |
| 🧰 App | Blur Blur Blur | Text to image generation with semantic control |
| 🏡 Onboarding | Starter Village | Guided entry point for new users |
If this repository helped, starring it improves discovery so more builders can find the docs and tools.
