mirror of
https://github.com/onestardao/WFGY.git
synced 2026-04-28 11:40:07 +00:00
| .. | ||
| blue_green_switchovers.md | ||
| cache_warmup_invalidation.md | ||
| db_migration_guardrails.md | ||
| feature_flags_safe_launch.md | ||
| idempotency_dedupe.md | ||
| incident_comms_and_statuspage.md | ||
| postmortem_and_regression_tests.md | ||
| rate_limit_backpressure.md | ||
| read_only_mode_and_maintenance_window.md | ||
| README.md | ||
| release_calendar_and_change_freeze.md | ||
| retry_backoff.md | ||
| rollback_and_fast_recovery.md | ||
| rollout_readiness_gate.md | ||
| shadow_traffic_mirroring.md | ||
| staged_rollout_canary.md | ||
| vector_index_build_and_swap.md | ||
| version_pinning_and_model_lock.md | ||
Ops & Deploy — Global Fix Map
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.
Open these first
- Visual map and recovery: RAG Architecture & Recovery
- End-to-end retrieval knobs: Retrieval Playbook
- Traceability and snippet schema: Retrieval Traceability · Data Contracts
- Boot order and deploy traps: Bootstrap Ordering · Deployment Deadlock · Pre-Deploy Collapse
- Live ops: Live Monitoring for RAG · Debug Playbook
When to use this folder
- First call after deploy crashes or returns 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 the old sections
- Retries cause duplicate side effects or money leaks
- Feature flags bleed traffic into unfinished paths
- Maintenance or migration windows corrupt anchors
Acceptance targets for a safe rollout
- ΔS(question, retrieved) ≤ 0.45 on three paraphrases after rollout begins
- Coverage ≥ 0.70 to the expected section on the new version
- λ remains convergent across two seeds during canary and after cutover
- Idempotency ≥ 99.9% on side-effecting actions during retry storms
- Zero silent index mismatches after build-and-swap (hash and doc counts match)
- P95 latency within budget while rate-limit backpressure is active
Quick routes to per-page guides
- Rollout gate: rollout_readiness_gate.md
- Canary playbook: staged_rollout_canary.md
- Blue/green cutover: blue_green_switchovers.md
- Version freeze: version_pinning_and_model_lock.md
- Vector index build-and-swap: vector_index_build_and_swap.md
- Cache warmup and invalidation: cache_warmup_invalidation.md
- Rate limit and backpressure: rate_limit_backpressure.md
- Feature flags, safe launch: feature_flags_safe_launch.md
- Idempotency and dedupe: idempotency_dedupe.md
- Retry policies and backoff: retry_backoff.md
- Rollback and fast recovery: rollback_and_fast_recovery.md
- Postmortem and regression tests: postmortem_and_regression_tests.md
- Release calendar and change freeze: release_calendar_and_change_freeze.md
- Incident comms and Statuspage: incident_comms_and_statuspage.md
- Shadow traffic mirroring: shadow_traffic_mirroring.md
- Read-only mode and maintenance: read_only_mode_and_maintenance_window.md
- DB migration guardrails: db_migration_guardrails.md
60-second ship checklist
-
Freeze the world
Pin model IDs, tool schemas, prompt headers, retriever weights. RecordMODEL_LOCK,PROMPT_REV,INDEX_HASH. -
Warm up safely
Build new index off-path. Validate counts, hashes, sample ΔS. Preload caches with canary questions. -
Shadow then canary
Mirror production queries. Compare ΔS, coverage, latency, and error rate. If stable, release to 5% then 25% then 100%. -
Guard the edge
Turn on backpressure at the gateway. Enable idempotency keys and bounded retries before any side effects. -
Know your exit
Prepare rollback switch and comms template. Keep Statuspage draft and regression pack ready.
Symptoms → exact fix
| What you see | Open this |
|---|---|
| First requests after deploy hit old snippets | vector_index_build_and_swap.md · cache_warmup_invalidation.md |
| Canary clean, full rollout flips answers | staged_rollout_canary.md · feature_flags_safe_launch.md |
| Wrong model responds after failover | version_pinning_and_model_lock.md |
| Retries cause duplicate charges or writes | idempotency_dedupe.md · retry_backoff.md |
| RL storms, queue explosion, timeouts | rate_limit_backpressure.md |
| Need instant safe rollback plan | rollback_and_fast_recovery.md · blue_green_switchovers.md |
| Maintenance window corrupts anchors | read_only_mode_and_maintenance_window.md · db_migration_guardrails.md |
| Unsure if it is safe to ship | rollout_readiness_gate.md |
🔗 Quick-Start Downloads (60 sec)
| 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 |
🧭 Explore More
| Module | Description | Link |
|---|---|---|
| WFGY Core | WFGY 2.0 engine is live: full symbolic reasoning architecture and math stack | View → |
| Problem Map 1.0 | Initial 16-mode diagnostic and symbolic fix framework | View → |
| Problem Map 2.0 | RAG-focused failure tree, modular fixes, and pipelines | View → |
| Semantic Clinic Index | Expanded failure catalog: prompt injection, memory bugs, logic drift | View → |
| Semantic Blueprint | Layer-based symbolic reasoning & semantic modulations | View → |
| Benchmark vs GPT-5 | Stress test GPT-5 with full WFGY reasoning suite | View → |
| 🧙♂️ Starter Village 🏡 | New here? Lost in symbols? Click here and let the wizard guide you through | Start → |
👑 Early Stargazers: See the Hall of Fame —
Engineers, hackers, and open source builders who supported WFGY from day one.
⭐ WFGY Engine 2.0 is already unlocked. ⭐ Star the repo to help others discover it and unlock more on the Unlock Board.
