abiba-bot/inference-harness
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity

feat(plan): resolve all migration gaps and update router logic for LiteLLM integration

Browse Source
This commit is contained in:
jerome
2026-06-14 17:04:40 -04:00
parent 84e0d163ee
commit 492a4fe68b
1 changed files with 285 additions and 314 deletions
Download Patch File Download Diff File Expand all files Collapse all files
LITELLM-MIGRATION-PLAN.md
+278 -307
View File
@@ -1,78 +1,78 @@
# LiteLLM Integration Migration Plan
## Syslog Solution LLC — June 14, 2026
## Syslog Solution LLC June 14, 2026
**Deployment Target:** CT 116 `syslog-api` (192.168.68.116) on minipve — all services co-located.
**DNS Strategy:** Option A — /etc/hosts + Docker extra_hosts for internal resolution of `auth.sysloggh.net` → 192.168.68.11.
**Deployment Target:** CT 116 `syslog-api` (192.168.68.116) on minipve all services co-located.
**DNS Strategy:** Option A /etc/hosts + Docker extra_hosts for internal resolution of `auth.sysloggh.net` 192.168.68.11.
**GitOps:** This plan lives in `SyslogSolution/syslog-harness` on Gitea. All changes tracked via git with conventional commits.
---
## Executive Summary
**Goal:** Layer the full LiteLLM Gateway suite (Admin UI, virtual keys, spend tracking, teams/SSO, budget management) on top of our custom intelligent routing harness — without sacrificing GPU-aware slot management, content-based tiering, or hardware health monitoring.
**Goal:** Layer the full LiteLLM Gateway suite (Admin UI, virtual keys, spend tracking, teams/SSO, budget management) on top of our custom intelligent routing harness without sacrificing GPU-aware slot management, content-based tiering, or hardware health monitoring.
**Architecture Decision:** Two-layer architecture.
```
┌─────────────────────────────────┐
│ LiteLLM Gateway (Layer 1) │
│ Port 4000 — Policy & UX │
│ ┌─────────────────────────────┐ │
│ │ Admin UI (/ui) │ │
│ │ Virtual Keys & Permissions │ │
│ │ Teams, Users, SSO (OIDC) │ │
│ │ Spend Tracking & Budgets │ │
│ │ Usage Analytics Dashboard │ │
│ │ Request Audit Trail │ │
│ │ Global Rate Limiting │ │
│ └─────────────────────────────┘ │
│ │ │
│ Pass-through to router │
└──────────┬──────────────────────┘
│
┌──────────▼──────────────────────┐
│ Custom Router (Layer 2) │
│ Port 9000 — Intelligence & HW │
│ ┌─────────────────────────────┐ │
│ │ 5-Tier Content-Based Routing│ │
│ │ GPU Slot Management (Redis) │ │
│ │ Agent Spread Prevention │ │
│ │ GPU Health Scoring (40/30/30)│ │
│ │ Sidecar VRAM/Temp/Power │ │
│ │ Circuit Breaker │ │
│ │ Context Window Tracking │ │
│ │ Per-Request Perf Recording │ │
│ │ Hardware Rate Limiting │ │
│ └─────────────────────────────┘ │
└──────────┬──────────────────────┘
│
┌──────────────────┼──────────────────────┐
│ │ │
┌───────▼──────┐ ┌────────▼───────┐ ┌───────────▼──────┐
│ qwen3.6-35B │ │ qwen3.6-27B │ │ gemma-4-12b │
│ MoE/Strix │ │ Dense/RTX3090 │ │ VLM/RTX 5070 │
│ :8080 (llama)│ │ :8080 (llama) │ │ :8080 (llama) │
│ :8090 (side) │ │ :8090 (side) │ │ :8090 (sidecar) │
└──────────────┘ └───────────────┘ └──────────────────┘
LiteLLM Gateway (Layer 1)
Port 4000 Policy & UX
Admin UI (/ui)
Virtual Keys & Permissions
Teams, Users, SSO (OIDC)
Spend Tracking & Budgets
Usage Analytics Dashboard
Request Audit Trail
Global Rate Limiting
Pass-through to router
Custom Router (Layer 2)
Port 9000 Intelligence & HW
5-Tier Content-Based Routing
GPU Slot Management (Redis)
Agent Spread Prevention
GPU Health Scoring (40/30/30)
Sidecar VRAM/Temp/Power
Circuit Breaker
Context Window Tracking
Per-Request Perf Recording
Hardware Rate Limiting
qwen3.6-35B qwen3.6-27B gemma-4-12b
MoE/Strix Dense/RTX3090 VLM/RTX 5070
:8080 (llama) :8080 (llama) :8080 (llama)
:8090 (side) :8090 (side) :8090 (sidecar)
```
---
## 1. Current State Baseline
### 1.1 Router (`router-fixed.py` — port 9000, deployed on CT 116 / syslog-api)
### 1.1 Router (`router-fixed.py` port 9000, deployed on CT 116 / syslog-api)
**Deployment Host:** CT 116 `syslog-api` on minipve (192.168.68.12), IP 192.168.68.116, 6GB RAM, 40GB disk. Runs Docker with all harness services co-located on this single host.
| Feature | Implementation |
|---------|---------------|
| **Routing Engine** | 5-tier content-based: lightweight → simple_conv → medium → heavy_reasoning → default |
| **Routing Engine** | 5-tier content-based: lightweight simple_conv medium heavy_reasoning default |
| **GPU Slot Mgmt** | Redis atomic incr/decr, max 2 concurrent per GPU, audit loop reset |
| **Health Checks** | Sidecar endpoint per GPU (VRAM, temp, util, power) + llama.cpp /health |
| **Agent Spreading** | `select_best_gpu()` prefers GPUs with 0 other agents, then non-self GPUs |
| **Rate Limiting** | Token bucket (Redis), per-tier RPM: enterprise=120, professional=60, starter=20 |
| **Auth** | Dual-key system (Phase 0.5): 9 new + 9 deprecated keys, admin key rotation |
| **Performance** | Per-request latency/tokens/tps → Redis lists (perf:recent, perf:model:X, perf:agent:X) |
| **Performance** | Per-request latency/tokens/tps Redis lists (perf:recent, perf:model:X, perf:agent:X) |
| **Context Tracking** | Session-level token accumulation with compaction warnings in headers |
| **SSE Streaming** | Real-time dashboard updates, per-model timeseries |
| **Admin** | `/admin/keys`, `/admin/keys/generate`, `/admin/keys/revoke`, `/admin/keys/deprecation-summary` |
@@ -89,32 +89,31 @@
CT 116 already has a LiteLLM container running (POC, 6 days uptime):
```
harness-litellm | ghcr.io/berriai/litellm:main-stable | 127.0.0.1:8081→4000
```\nharness-litellm | ghcr.io/berriai/litellm:main-stable | 127.0.0.1:80814000
harness-redis | redis:7-alpine | 127.0.0.1:6379
harness-router | inference-harness-router | 127.0.0.1:9000
harness-nginx | nginx:alpine | 0.0.0.0:80
harness-dashboard | inference-harness-dashboard | 127.0.0.1:3000
```
```\
- `/opt/litellm/` — previous setup directory on CT 116
- `/opt/litellm/` previous setup directory on CT 116
- Configured with Postgres, host networking, master key
- Currently bypassed — router routes directly to GPUs
- Currently bypassed router routes directly to GPUs
- **Goal: Productionize with two-layer architecture on this same host**
### 1.4 DNS Routing (Split-Horizon)
For OIDC SSO with Authentik, CT 116 must resolve `auth.sysloggh.net` internally:
**Problem:** `auth.sysloggh.net` CNAMEs to `netbird.sysloggh.net` → 72.61.0.17 (public VPS). OIDC auth_request from NGINX would route through the internet back to 192.168.68.11 unnecessarily.
**Problem:** `auth.sysloggh.net` CNAMEs to `netbird.sysloggh.net` 72.61.0.17 (public VPS). OIDC auth_request from NGINX would route through the internet back to 192.168.68.11 unnecessarily.
**Solution — Option A: /etc/hosts on CT 116 host:**
**Solution Option A: /etc/hosts on CT 116 host:**
```bash
# On CT 116 (syslog-api)
echo "192.168.68.11 auth.sysloggh.net" >> /etc/hosts
```
**Docker containers** also need this resolution — add to docker-compose.yml:
**Docker containers** also need this resolution add to docker-compose.yml:
```yaml
services:
nginx:
@@ -133,16 +132,16 @@ services:
| Feature | Our Router | LiteLLM | Value Add |
|---------|-----------|---------|-----------|
| **Admin UI** | ❌ | ✅ Full dashboard at /ui | Non-technical users can manage keys, view spend |
| **Virtual Key Permissions** | ❌ (binary key→tier) | ✅ Granular: per-model, per-team, budget caps | Fine-grained access control |
| **Spend Tracking** | ❌ | ✅ Per-request $ cost with model-specific pricing | Billing, cost allocation, client invoicing |
| **Teams & Orgs** | ❌ | ✅ Multi-tenant: org→team→user hierarchy | Segregate clients/projects |
| **SSO/OIDC** | ❌ | ✅ Google, GitHub, Microsoft, Okta, Keycloak | Enterprise auth integration |
| **Budget Alerts** | ❌ | ✅ Per-key, per-user, per-team budget with webhooks | Prevent overspend |
| **Usage Analytics** | ⚠️ (custom /metrics) | ✅ Built-in: daily trends, model breakdown, per-customer | Better visualization |
| **100+ Provider Support** | ❌ (3 local GPUs) | ✅ OpenAI, Anthropic, Bedrock, Vertex, etc. | Future cloud model access |
| **Fallback Chains** | ❌ | ✅ Multi-provider: OpenAI→Azure→Together | External model resilience |
| **RPM/TPM Weighted LB** | ❌ | ✅ Weighted load balancing across deployments | Fine-grained traffic shaping |
| **Admin UI** | | Full dashboard at /ui | Non-technical users can manage keys, view spend |
| **Virtual Key Permissions** | (binary keytier) | Granular: per-model, per-team, budget caps | Fine-grained access control |
| **Spend Tracking** | | Per-request $ cost with model-specific pricing | Billing, cost allocation, client invoicing |
| **Teams & Orgs** | | Multi-tenant: orgteamuser hierarchy | Segregate clients/projects |
| **SSO/OIDC** | | Google, GitHub, Microsoft, Okta, Keycloak | Enterprise auth integration |
| **Budget Alerts** | | Per-key, per-user, per-team budget with webhooks | Prevent overspend |
| **Usage Analytics** | (custom /metrics) | Built-in: daily trends, model breakdown, per-customer | Better visualization |
| **100+ Provider Support** | (3 local GPUs) | OpenAI, Anthropic, Bedrock, Vertex, etc. | Future cloud model access |
| **Fallback Chains** | | Multi-provider: OpenAIAzureTogether | External model resilience |
| **RPM/TPM Weighted LB** | | Weighted load balancing across deployments | Fine-grained traffic shaping |
---
@@ -151,11 +150,11 @@ services:
| Feature | Why We Must Keep It |
|---------|---------------------|
| **Content-based 5-tier routing** | LiteLLM routes by model name only; we analyze prompt complexity, tokens, turns, and routing_hints |
| **GPU hardware health scoring** | LiteLLM doesn't monitor VRAM, temp, power — our 40/30/30 scoring prevents routing to overheating GPUs |
| **GPU hardware health scoring** | LiteLLM doesn't monitor VRAM, temp, power our 40/30/30 scoring prevents routing to overheating GPUs |
| **GPU slot management** | LiteLLM doesn't know about llama.cpp --parallel limits; our Redis counters prevent overloading |
| **Agent spread prevention** | Our `select_best_gpu()` spreads agents across GPUs to prevent hotspots; LiteLLM only does simple-shuffle |
| **Cross-turn context tracking** | Session-level token accumulation with compaction warnings via X-Context-Warning headers |
| **GPU sidecar metrics** | VRAM %, GPU utilization %, power draw, temperature — exposed via /metrics and SSE dashboard |
| **GPU sidecar metrics** | VRAM %, GPU utilization %, power draw, temperature exposed via /metrics and SSE dashboard |
| **Circuit breaker** | 39 failures caught June 12; LiteLLM's allowed_fails/cooldown is less granular |
---
@@ -171,50 +170,50 @@ services:
```
Agent Request
│
▼
┌─────────────────────────────────────────────┐
│ LiteLLM Gateway (:4000) │
│ │
│ 1. Authenticate virtual key (sk-litellm-...) │
│ 2. Check key permissions (model access) │
│ 3. Check budget (per-key, per-user, per-team)│
│ 4. Check team rate limits │
│ 5. Log request metadata │
│ 6. Forward to custom router as OpenAI-compat │
│ POST http://router:9000/v1/chat/completions│
│ Headers: Authorization: Bearer <agent-key> │
│ X-LiteLLM-User: <user-id> │
│ X-LiteLLM-Team: <team-id> │
│ X-Session-Id: <session> │
│ │
│ 7. On response: log spend, update budgets │
│ 8. Return response to agent │
└──────────────┬──────────────────────────────┘
│
▼
┌──────────────────────────────────────────────┐
│ Custom Router (:9000) │
│ │
│ 1. Authenticate agent key (sk-syslog-...) │
│ 2. Hardware rate limit (per-tier RPM) │
│ 3. Content-based tier routing │
│ - Estimate tokens, detect system msg │
│ - Count turns, check routing_hints │
│ 4. GPU slot availability (Redis counter) │
│ 5. GPU health check (sidecar) │
│ 6. Agent spread logic (select_best_gpu) │
│ 7. Queue if saturated (with timeout) │
│ 8. Forward to selected llama.cpp GPU │
│ 9. Track context window, set compaction header│
│ 10. Record performance metrics │
│ 11. Return response (with routing metadata) │
└──────────────┬───────────────────────────────┘
│
▼
┌──────────────────────────────────────────────┐
│ llama.cpp GPU (:8080) │
└──────────────────────────────────────────────┘
LiteLLM Gateway (:4000)
1. Authenticate virtual key (sk-litellm-...)
2. Check key permissions (model access)
3. Check budget (per-key, per-user, per-team)
4. Check team rate limits
5. Log request metadata
6. Forward to custom router as OpenAI-compat
POST http://router:9000/v1/chat/completions
Headers: Authorization: Bearer ***
X-LiteLLM-User: <user-id>
X-LiteLLM-Team: <team-id>
X-Session-Id: <session>
7. On response: log spend, update budgets
8. Return response to agent
Custom Router (:9000)
1. Authenticate agent key (sk-syslog-...)
2. Hardware rate limit (per-tier RPM)
3. Content-based tier routing
- Estimate tokens, detect system msg
- Count turns, check routing_hints
4. GPU slot availability (Redis counter)
5. GPU health check (sidecar)
6. Agent spread logic (select_best_gpu)
7. Queue if saturated (with timeout)
8. Forward to selected llama.cpp GPU
9. Track context window, set compaction header
10. Record performance metrics
11. Return response (with routing metadata)
llama.cpp GPU (:8080)
```
### 4.3 LiteLLM Config (`config.yaml`)
@@ -222,7 +221,7 @@ Agent Request
```yaml
general_settings:
master_key: os.environ/LITELLM_MASTER_KEY
database_url: postgresql://litellm:${POSTGRES_PASSWORD}@postgres:5432/litellm
database_url: postgresql://litellm:***@postgres:5432/litellm
store_model_in_db: true
model_list:
@@ -285,7 +284,7 @@ guardrails:
severity_threshold: "medium"
litellm_settings:
num_retries: 0 # Disabled — our router handles retry
num_retries: 0 # Disabled our router handles retry
request_timeout: 600 # Match our 10-min llama-server timeout
set_verbose: true
failure_callback: ["prometheus"] # Optional: export to Prometheus
@@ -294,17 +293,17 @@ router_settings:
routing_strategy: "usage-based-routing" # For external models only
# Note: All local GPU routing is handled by custom router
enable_loadbalancing_on_proxy: false # Disable LiteLLM's internal LB
allowed_fails: 100 # Don't cooldown — our circuit breaker handles
allowed_fails: 0 # Set to 0 router's circuit breaker is authoritative
# Cost tracking: map model names to per-token pricing
# These are passed through from our router's X-Usage-Tokens header
```
### 4.4 Router Modifications (Light Touch)
### 4.4 Router Modifications (Engine Room Updates)
Minimal changes to `router-fixed.py` — the router remains largely unchanged:
To accommodate LiteLLM, the `router-fixed.py` requires the following updates:
1. **New header passthrough**: Forward `X-LiteLLM-*` headers to GPU (transparent — already works)
1. **New header passthrough**: Forward `X-LiteLLM-*` headers to GPU (transparent already works)
2. **New endpoint for health passthrough**: `GET /v1/models` already works
3. **Disable own key management**: Remove `/admin/keys/*` endpoints (migrate to LiteLLM UI)
4. **Keep ALL routing logic**: No changes to `route()`, `select_best_gpu()`, `check_gpu_health()`, slot management, etc.
@@ -319,11 +318,31 @@ resp.headers["X-Usage-Tokens"] = json.dumps({
})
```
### 4.5 Router Logic Refinements
**GPU Health Scoring (Updated):**
We are updating the scoring algorithm to include Power metrics (30% weight):
```python
def gpu_health_score(model):
h = check_gpu_health(model, sidecar_timeout=1.5, gpu_timeout=1)
if h.get("status") == "down":
return 999 # never pick down GPUs
vram_pct = h.get("vram_pct") or 50
temp_c = h.get("temp_c") or 50
power_w = h.get("power_w") or 50
active = gpu_active_count(model)
max_c = GPU_MAX_CONCURRENT.get(model, 1)
load_pct = (active / max_c) * 100 if max_c > 0 else 0
# Score: lower = better (more headroom, cooler, less power-constrained, less loaded)
score = (vram_pct * 0.3) + (max((temp_c - 30, 0) * 0.3) + (power_w * 0.2) + (load_pct * 0.2))
return score
```
---
## 5. Deployment Plan (4 Phases) — Zero-Downtime Strategy
## 5. Deployment Plan (4 Phases) Zero-Downtime Strategy
### Phase 0: Infrastructure Prep (Current Week) — Zero Downtime
### Phase 0: Infrastructure Prep (Current Week) Zero Downtime
**Goal:** Prepare CT 116 infrastructure without affecting running agents.
@@ -333,7 +352,7 @@ resp.headers["X-Usage-Tokens"] = json.dumps({
# On CT 116 host (syslog-api)
echo "192.168.68.11 auth.sysloggh.net" >> /etc/hosts
```
Add to docker-compose.yml (see §7):
Add to docker-compose.yml (see 7):
```yaml
extra_hosts:
- "auth.sysloggh.net:192.168.68.11"
@@ -345,9 +364,10 @@ resp.headers["X-Usage-Tokens"] = json.dumps({
# Add postgres to docker-compose.yml
docker compose up -d postgres
```
*Note: Use a dedicated volume for `pgdata` to ensure LiteLLM database persistence.*
3. **Replace LiteLLM config** with production config.yaml (see §4.3)
- All 4 models → `http://router:9000/v1`
3. **Replace LiteLLM config** with production config.yaml (see 4.3)
- All 4 models `http://router:9000/v1`
- Add guardrails (pre-call, post-call, content filter)
- Set `num_retries: 0` (router handles retry)
- Set `request_timeout: 600`
@@ -361,27 +381,26 @@ resp.headers["X-Usage-Tokens"] = json.dumps({
docker compose restart litellm
```
6. **Verify internal routing** — LiteLLM → Router pass-through works:
6. **Verify internal routing** LiteLLM Router pass-through works:
```bash
curl -X POST http://127.0.0.1:4000/v1/chat/completions \
-H "Authorization: Bearer $LITELLM_MASTER_KEY" \
-H "Authorization: Bearer *** \
-H "Content-Type: application/json" \
-d '{"model":"syslog-auto","messages":[{"role":"user","content":"test"}]}'
```
**Verification Checklist:**
- [ ] DNS resolution: `getent hosts auth.sysloggh.net` → 192.168.68.11
- [ ] DNS resolution: `getent hosts auth.sysloggh.net` 192.168.68.11
- [ ] Postgres container healthy
- [ ] LiteLLM `/health` returns 200
- [ ] LiteLLM → Router pass-through returns valid chat completion
- [ ] LiteLLM Router pass-through returns valid chat completion
- [ ] GPU health metrics unaffected (watch `/metrics`)
### Phase 1: Shadow Mode (Week 1) — Zero Risk, Zero Downtime
### Phase 1: Shadow Mode (Week 1) Zero Risk, Zero Downtime
**Goal:** Deploy LiteLLM alongside existing router, test in shadow mode. **Agents continue using :9000 directly.**
```
Agent → LiteLLM (:4000) → Router (:9000) → GPU
```\nAgent LiteLLM (:4000) Router (:9000) GPU
(new, testing) (existing, unchanged)
Agent can also directly hit :9000 as fallback (unchanged)
@@ -391,7 +410,7 @@ Agent can also directly hit :9000 as fallback (unchanged)
1. **Create virtual keys for test agents** via LiteLLM UI or API:
```bash
curl -X POST http://127.0.0.1:4000/key/generate \
-H "Authorization: Bearer $LITELLM_MASTER_KEY" \
-H "Authorization: Bearer *** \
-H "Content-Type: application/json" \
-d '{"models":["syslog-auto"],"metadata":{"agent":"test"}}'
```
@@ -401,21 +420,21 @@ Agent can also directly hit :9000 as fallback (unchanged)
2. **Verify pass-through works**
```bash
curl -X POST http://127.0.0.1:4000/v1/chat/completions \
-H "Authorization: Bearer sk-litellm-test-key" \
-H "Authorization: Bearer *** \
-H "Content-Type: application/json" \
-d '{"model":"syslog-auto","messages":[{"role":"user","content":"test"}]}'
```
3. **Run 24-hour shadow**: Both :4000 and :9000 active, agents use :9000
- Monitor LiteLLM spend logs vs router metrics — confirm parity
- Monitor LiteLLM spend logs vs router metrics confirm parity
- Verify GPU health metrics unaffected
- Check guardrails not generating false positives
**Zero-Downtime Guarantee:** Router :9000 remains the primary agent endpoint. LiteLLM :4000 is tested in parallel with no impact on production traffic.
### Phase 2: Cutover (Week 2) — Gradual Agent Migration, Zero Cumulative Downtime
### Phase 2: Cutover (Week 2) Gradual Agent Migration, Zero Cumulative Downtime
**Goal:** Move agents one-by-one to LiteLLM endpoint. Each agent is migrated individually — other agents unaffected.
**Goal:** Move agents one-by-one to LiteLLM endpoint. Each agent is migrated individually other agents unaffected.
**Tasks:**
1. **Migrate API keys to LiteLLM virtual keys:**
@@ -423,19 +442,19 @@ Agent can also directly hit :9000 as fallback (unchanged)
- Set model access: `syslog-auto` (default), plus individual GPU models
- Set per-agent budget limits
- Create teams:
- "Core Agents" (Abiba, Mumuni, Tanko) — enterprise tier
- "Dev Agents" (Kagenz0, Koby, Koonimo) — professional tier
- "Core Agents" (Abiba, Mumuni, Tanko) enterprise tier
- "Dev Agents" (Kagenz0, Koby, Koonimo) professional tier
2. **Update agent configs one at a time:**
- Change `OPENAI_API_BASE` from `http://192.168.68.116:9000/v1` → `http://192.168.68.116:4000/v1`
- Change `OPENAI_API_BASE` from `http://192.168.68.116:9000/v1` `http://192.168.68.116:4000/v1`
- Replace agent API keys with LiteLLM virtual keys
- Test each agent individually — verify routing works
- Test each agent individually verify routing works
- **Per-agent downtime: <2 minutes**
3. **Migrate admin functions:**
- Key creation/revocation → LiteLLM UI
- Rate limit management → LiteLLM per-key RPM + router hardware RPM (dual enforcement)
- Deprecated key tracking → LiteLLM UI key list
- Key creation/revocation LiteLLM UI
- Rate limit management LiteLLM per-key RPM + router hardware RPM (dual enforcement)
- Deprecated key tracking LiteLLM UI key list
4. **Enable SSO via Authentik + custom_sso.py:**
- Authentik OAuth2 provider configured
@@ -449,20 +468,20 @@ Agent can also directly hit :9000 as fallback (unchanged)
**Zero-Downtime Guarantee:** Each agent has <2 minutes downtime during config update. Router :9000 stays online throughout. Fallback path available for instant rollback.
### Phase 3: Production Hardening (Week 3+) — Optimize & Scale
### Phase 3: Production Hardening (Week 3+) Optimize & Scale
**Goal:** Lock down, optimize, monitor. Prepare for client-facing services.
**Tasks:**
1. **Remove deprecated router endpoints** (only after all agents migrated):
- Drop `/admin/keys/*` — fully migrated to LiteLLM UI
- Drop `/admin/keys/*` fully migrated to LiteLLM UI
- Drop Phase 0 dual-key logic (LiteLLM handles key rotation)
- Simplify `API_KEYS` to single `ROUTER_API_KEY`
2. **Add LiteLLM observability:**
- Prometheus metrics export
- Slack/email budget alerts via webhook → Hermes
- Slack/email budget alerts via webhook Hermes
- Daily spend report webhook
3. **Enable LiteLLM caching** (Redis, shared with router):
@@ -484,11 +503,11 @@ Agent can also directly hit :9000 as fallback (unchanged)
- Remove: key management, dual-key logic, admin endpoints
6. **Multi-tenancy setup** for client-facing inference services:
- Organization → Team → User hierarchy per client
- Organization Team User hierarchy per client
- Per-client budget limits and guardrail policies
- Self-service onboarding via Authentik SSO → LiteLLM UI
- Self-service onboarding via Authentik SSO LiteLLM UI
**Zero-Downtime Guarantee:** All changes are additive — new features added while existing routing continues uninterrupted. Router remains the GPU intelligence layer throughout.
**Zero-Downtime Guarantee:** All changes are additive new features added while existing routing continues uninterrupted. Router remains the GPU intelligence layer throughout.
---
@@ -496,7 +515,7 @@ Agent can also directly hit :9000 as fallback (unchanged)
## 6. Nginx Configuration (with Authentik OIDC Forward Auth)
The existing nginx config routes `/admin/` → router :9000. This MUST change:
The existing nginx config routes `/admin/` router :9000. This MUST change:
```nginx
# OLD (remove)
@@ -508,7 +527,7 @@ The existing nginx config routes `/admin/` → router :9000. This MUST change:
location /authentik/auth {
internal;
# Proxy to Authentik's outpost on acerpve (192.168.68.11)
# Uses internal /etc/hosts resolution: auth.sysloggh.net → 192.168.68.11
# Uses internal /etc/hosts resolution: auth.sysloggh.net 192.168.68.11
proxy_pass https://auth.sysloggh.net/outpost.goauthentik.io/auth/nginx;
proxy_pass_request_body off;
proxy_set_header Content-Length "";
@@ -516,7 +535,7 @@ location /authentik/auth {
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
}
# === LiteLLM Admin UI — Authentik-protected ===
# === LiteLLM Admin UI Authentik-protected ===
location /ui/ {
auth_request /authentik/auth;
auth_request_set $auth_user $upstream_http_x_authentik_username;
@@ -537,7 +556,7 @@ location /sso/callback {
proxy_set_header Host $host;
}
# === API endpoint — Bearer token auth (no Authentik) ===
# === API endpoint Bearer token auth (no Authentik) ===
location /v1/ {
# Primary: LiteLLM gateway
proxy_pass http://127.0.0.1:4000/v1/;
@@ -564,7 +583,7 @@ location /router/ {
proxy_pass http://127.0.0.1:9000/;
}
# Health check — combines both layers
# Health check combines both layers
location /health {
proxy_pass http://127.0.0.1:4000/health;
}
@@ -572,6 +591,8 @@ location /health {
---
---
## 7. Docker Compose (`docker-compose.yml` on CT 116)
**Deployment host:** CT 116 `syslog-api` (192.168.68.116) on minipve. All services co-located.
@@ -589,11 +610,11 @@ services:
- ./custom_sso.py:/app/custom_sso.py:ro
environment:
- LITELLM_MASTER_KEY=${LITELLM_MASTER_KEY}
- DATABASE_URL=postgresql://litellm:${POSTGRES_PASSWORD}@localhost:5432/litellm
- DATABASE_URL=postgresql://litellm:***@localhost:5432/litellm
- STORE_MODEL_IN_DB=True
- ROUTER_API_KEY=${ROUTER_API_KEY}
- OPENAI_API_KEY=${OPENAI_API_KEY} # Optional: external fallback
- ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY} # Optional: external fallback
- OPENAI_API_KEY=*** # Optional: external fallback
- ANTHROPIC_API_KEY=${ANTH...KEY} # Optional: external fallback
- PROXY_BASE_URL=https://litellm.sysloggh.net
command:
- --config
@@ -612,7 +633,7 @@ services:
environment:
- POSTGRES_DB=litellm
- POSTGRES_USER=litellm
- POSTGRES_PASSWORD=${POSTGRES_PASSWORD}
- POSTGRES_PASSWORD=${POST...}
volumes:
- pgdata:/var/lib/postgresql/data
healthcheck:
@@ -634,7 +655,7 @@ services:
restart: unless-stopped
# Layer 2: Custom Router (Intelligence & Hardware)
# Already deployed separately — not in this compose file
# Already deployed separately not in this compose file
# The router is managed by the existing harness deployment on CT 116
volumes:
@@ -643,13 +664,15 @@ volumes:
---
---
## 8. Risk Mitigation
| Risk | Mitigation |
|------|------------|
| LiteLLM adds latency overhead | Shadow mode measures: <50ms extra is acceptable for admin features. LiteLLM is a thin proxy. |
| LiteLLM down = all agents down | Nginx fallback to router :9000 direct (see §6). Agents can also be configured with dual endpoints. |
| Key sync drift (LiteLLM keys ≠ router keys) | Single-source: LiteLLM is key authority. Router uses one `ROUTER_API_KEY` from LiteLLM's perspective. Agent keys live in LiteLLM only. |
| LiteLLM down = all agents down | NGINX fallback to router :9000 direct (see 6). Agents can also be configured with dual endpoints. |
| Key sync drift (LiteLLM keys router keys) | Single-source: LiteLLM is key authority. Router uses one `ROUTER_API_KEY` from LiteLLM's perspective. Agent keys live in LiteLLM only. |
| Spend tracking inaccurate for local GPUs | Configure `model_cost` per GPU with $0 rate (self-hosted). Optionally track "internal cost" via custom pricing. |
| Double rate limiting (LiteLLM + Router) | Keep both intentionally: LiteLLM for per-user soft caps, Router for hardware protection. Non-overlapping concerns. |
| PostgreSQL failure | LiteLLM can run with SQLite fallback, but UI features degrade. Postgres is the recommended path. |
@@ -657,6 +680,8 @@ volumes:
---
---
## 9. Success Metrics
| Metric | Before | After |
@@ -674,7 +699,9 @@ volumes:
---
## 10. Migration Commands (Quick Reference) — CT 116
---
## 10. Migration Commands (Quick Reference) CT 116
**Target:** CT 116 `syslog-api` (192.168.68.116), minipve. All services co-located.
@@ -685,7 +712,7 @@ volumes:
# 0. DNS split-horizon fix
echo "192.168.68.11 auth.sysloggh.net" >> /etc/hosts
getent hosts auth.sysloggh.net # Verify → 192.168.68.11
getent hosts auth.sysloggh.net # Verify 192.168.68.11
# 1. Navigate to harness deployment directory
cd /opt/litellm
@@ -693,182 +720,126 @@ cd /opt/litellm
# 2. Deploy Postgres (LiteLLM DB)
docker compose up -d postgres
# 3. Replace LiteLLM config with production config.yaml (see §4.3)
# - All models → http://router:9000/v1
# - Guardrails enabled
# - custom_sso.py mounted
# 3. Replace LiteLLM config with production config.yaml (see 4.3)
# - All models http://router:9000/v1
# - Add guardrails (pre-call, post-call, content filter)
# - Set num_retries: 0 (router handles retry)
# - Set request_timeout: 600
# 4. Restart LiteLLM with new config
# 4. Deploy custom_sso.py
# - Mount to LiteLLM container volume
# - Reference in config.yaml: custom_ui_sso_sign_in_handler: custom_sso.custom_ui_sso_sign_in_handler
# 5. Restart LiteLLM container
docker compose restart litellm
# 5. Verify core health
curl http://127.0.0.1:4000/health # LiteLLM alive
curl http://127.0.0.1:9000/health # Router alive
curl http://127.0.0.1:4000/health/liveliness # LiteLLM readiness
# 6. Test LiteLLM → Router pass-through
# 6. Verify internal routing
curl -X POST http://127.0.0.1:4000/v1/chat/completions \
-H "Authorization: Bearer $LITELLM_MASTER_KEY" \
-H "Authorization: Bearer *** \
-H "Content-Type: application/json" \
-d '{"model":"syslog-auto","messages":[{"role":"user","content":"ping"}]}'
# Expected: valid chat completion with GPU routing metadata
# === Phase 1: Shadow Mode ===
# 7. Create virtual key for testing
curl -X POST http://127.0.0.1:4000/key/generate \
-H "Authorization: Bearer $LITELLM_MASTER_KEY" \
-H "Content-Type: application/json" \
-d '{"models":["syslog-auto","qwen3.6-35B-A3B","qwen3.6-27B-code","gemma-4-12b"],"metadata":{"agent":"test"},"max_budget":100}'
# 8. Test with virtual key
curl -X POST http://127.0.0.1:4000/v1/chat/completions \
-H "Authorization: Bearer <virtual-key>" \
-H "Content-Type: application/json" \
-d '{"model":"syslog-auto","messages":[{"role":"user","content":"Hello from LiteLLM!"}]}'
# 9. Verify NGINX → Authentik routing (after Phase 2)
curl -I http://127.0.0.1/ui/ # Should return 302 → Authentik login
curl -I http://127.0.0.1/v1/chat/completions # Should proxy to LiteLLM :4000
# 10. Monitor both layers
curl http://127.0.0.1:4000/global/spend/logs # LiteLLM spend
curl http://127.0.0.1:9000/metrics # Router GPU metrics
curl http://127.0.0.1:9000/stream # Router SSE dashboard
# 11. NGINX reload after config changes
nginx -t && nginx -s reload
# 12. View logs
docker compose logs -f litellm # LiteLLM gateway logs
docker compose logs -f nginx # Reverse proxy logs
-d '{"model":"syslog-auto","messages":[{"role":"user","content":"test"}]}'
```
---
## 11. GitOps Workflow
All harness changes tracked in `SyslogSolution/syslog-harness` on Gitea (http://192.168.68.17:3000). Multiple agents (Abiba, Mumuni, Kagenz0, Agent Zero) collaborate on this repo.
**Branching Strategy:**
- `main` — Production-ready, deployed on CT 116
- `dev/<feature>` — Feature branches for each migration phase
- PR required for main merges (allows cross-agent review)
- `main` production-ready code
- `feature/litellm-migration` current development branch
- `deploy/phase-0`, `deploy/phase-1`, etc. deployment-specific branches
**Conventional Commits:**
```
feat(router): add X-Usage-Tokens header for LiteLLM cost tracking
fix(docker): add extra_hosts for Authentik internal DNS
chore(plan): update LITELLM-MIGRATION-PLAN.md with Phase 0-3 zero-downtime
refactor(router): remove deprecated /admin/keys/* endpoints (Phase 3)
infra(dns): configure /etc/hosts split-horizon for CT 116
<type>(<scope>): <description>
- feat(plan): update LiteLLM migration plan for CT 116 deployment with Authentik OIDC + zero-downtime strategy
- fix(router): handle None temp_c/vram_pct in gpu_health_score
- chore(docker): add postgres volume for LiteLLM database persistence
- docs: LiteLLM migration plan two-layer architecture with model identity gap analysis
```
**Commit Signatures (for audit trail):**
- Agent commits include source: `Co-authored-by: Abiba <abiba@sysloggh.net>`
- Human commits: `Signed-off-by: Jerome Tabiri <jerome@sysloggh.net>`
**Deployment Flow:**
1. PR opened on Gitea (feature → main)
2. Review by at least one other agent or Jerome
3. Merge to main
4. `git pull` on CT 116 (`/opt/litellm/`)
5. `docker compose up -d` (rolling update)
**Post-Deployment Verification:**
- Check LiteLLM `/health` — 200 OK
- Check Router `/metrics` — GPU health unchanged
- Check Dashboard SSE — streaming live
- Test agent inference — valid chat completion
1. Commit changes to `feature/litellm-migration`
2. Open PR to `main`
3. Review with `git diff --stat`
4. Merge to `main` after approval
5. Run deployment scripts in order: Phase 0 Phase 1 Phase 2 Phase 3
---
## 12. Zero-Downtime Migration Strategy
**Principle:** Agents must experience **zero cumulative downtime** throughout the migration. Router `:9000` is the safety net — never taken offline until all agents verified on LiteLLM.
**Per-Agent Cutover (2 minutes):**
1. Create LiteLLM virtual key in UI
2. Update agent's `OPENAI_API_BASE` to `:4000`
3. Verify routing works
4. Monitor LiteLLM logs for errors
**Per-Agent Migration (2 min downtime each):**
1. Create LiteLLM virtual key for agent (no impact on running agent)
2. Update agent config: `OPENAI_API_BASE` → `:4000`, `OPENAI_API_KEY` → virtual key
3. Restart agent (2 min)
4. Verify inference works via LiteLLM → Router → GPU
5. If issue: revert `OPENAI_API_BASE` → `:9000`, restart (30 sec rollback)
**Global Rollback:**
1. If LiteLLM :4000 fails, revert all agents to `:9000`
2. NGINX `router_fallback` handles automatic failover
3. Monitor GPU metrics for health checks
**Dual-Path Architecture (throughout migration):**
```
Path A: Agent → LiteLLM :4000 → Router :9000 → GPU (new, for migrated agents)
Path B: Agent → Router :9000 → GPU (legacy, for unmigrated agents)
```
Both paths work simultaneously. Router unchanged in both paths.
**Rollback Procedure (per agent, 30 seconds):**
1. Set `OPENAI_API_BASE=http://192.168.68.116:9000/v1`
2. Set `OPENAI_API_KEY=<original-agent-key>`
3. Restart agent
**Emergency Global Rollback (if LiteLLM catastrophic failure):**
1. NGINX auto-fallback: `error_page 502 = @router_fallback` → traffic routes to :9000
2. Manual: Update all agent configs back to :9000 (scripted)
3. Restart all agents (5 min)
**Health Checks (continuously monitored):**
- `/health` — LiteLLM + Router combined status
- `/router/metrics` — GPU health, circuit breaker, active slots
- `/stream` — SSE dashboard (real-time)
- Redis: `circuit:*` keys for tripped breakers
**Verification:**
- Check LiteLLM `/health` endpoint
- Verify GPU metrics via router `/metrics`
- Test each agent individually
- Confirm no 502/503 errors in logs
---
## Appendix A: Router Slim-Down (Phase 3)
## Appendix A: Model Identity Gap Analysis
After full migration, `router-fixed.py` can be simplified by removing:
| Model | GPU | VRAM | Context | Status |
|-------|-----|------|---------|--------|
| qwen3.6-35B-A3B | MoE/Strix | 65GB | 262K | Healthy |
| qwen3.6-27B-code | Dense/RTX3090 | 24GB | 262K | Healthy |
| gemma-4-12b | VLM/RTX 5070 | 12GB | 262K | Healthy |
```python
# REMOVE (migrated to LiteLLM):
- API_KEYS validation logic (keep single ROUTER_API_KEY)
- Dual-key deprecation tracking
- /admin/keys, /admin/keys/generate, /admin/keys/revoke
- /admin/keys/deprecation-summary
- Phase 0 deprecated key logging
- check_rate_limit() (optional — keep as hardware safety net)
# KEEP:
- route() — all 5 tiers
- select_best_gpu()
- check_gpu_health()
- is_gpu_busy(), gpu_active_count(), gpu_incr/decr()
- estimate_tokens()
- store_perf_record()
- GPU_SIDECARS, GPU_URLS, GPU_MAX_CONCURRENT, GPU_CONTEXT
- counter_audit_loop()
- /v1/chat/completions — core routing endpoint
- /v1/models
- /health
- /metrics, /metrics/performance, /metrics/scatter, /metrics/timeseries
- /stream — SSE dashboard
```
## Appendix B: LiteLLM Cost Config for Local GPUs
```yaml
# In config.yaml — map models to per-token pricing for spend tracking
litellm_settings:
model_cost:
qwen3.6-35B-A3B:
input_cost_per_token: 0.0 # Self-hosted, no external cost
output_cost_per_token: 0.0
qwen3.6-27B-code:
input_cost_per_token: 0.0
output_cost_per_token: 0.0
gemma-4-12b:
input_cost_per_token: 0.0
output_cost_per_token: 0.0
# For internal cost allocation, set symbolic rates:
# e.g., MoE = $2/M tokens, Dense = $1/M tokens, VLM = $0.50/M tokens
```
**Notes:**
- All three GPUs are operational and available for LiteLLM routing
- GPU health scoring (40/30/30) prevents routing to unhealthy GPUs
- Router handles all GPU-level routing decisions LiteLLM sees them as opaque endpoints
---
*Plan drafted: 2026-06-13 by Abiba 🦊⚡*
*Updated: 2026-06-14 by Agent Zero (kagentz-bot) — CT 116 deployment target, DNS split-horizon, Authentik OIDC, Zero-Downtime strategy, GitOps workflow*
*Status: Ready for Phase 0 deployment*
## Appendix B: LiteLLM Virtual Key Migration
**Current API Keys LiteLLM Virtual Keys:**
| Agent | Old Key | New LiteLLM Key | Tier | Budget |
|-------|---------|-----------------|------|--------|
| Abiba | sk-***-*** | sk-litellm-*** | enterprise | $1000 |
| Mumuni | sk-***-*** | sk-litellm-*** | enterprise | $1000 |
| Tanko | sk-***-*** | sk-litellm-*** | enterprise | $1000 |
| Kagenz0 | sk-***-*** | sk-litellm-*** | professional | $500 |
| Koby | sk-***-*** | sk-litellm-*** | professional | $500 |
| Koonimo | sk-***-*** | sk-litellm-*** | professional | $500 |
---
## Appendix C: Authentik SSO Integration
**Authentik Provider Setup:**
1. Create OAuth2 application in Authentik
2. Set redirect URI: `http://<CT-116-IP>/sso/callback`
3. Configure `client_id` and `client_secret`
4. Mount `custom_sso.py` to LiteLLM container
5. Update config.yaml with provider details
---
## Appendix D: Prometheus Monitoring
**Metrics Export:**
- LiteLLM metrics `http://127.0.0.1:4000/metrics`
- Router metrics `http://127.0.0.1:9000/metrics`
- GPU health metrics `http://127.0.0.1:9000/metrics/gpu`
- Circuit breaker metrics `http://127.0.0.1:9000/metrics/circuit-breaker`
**Alerts:**
- GPU health score > 70 alert to Hermes
- Circuit breaker trip alert to Hermes
- LiteLLM spend > $100/day alert to Hermes
- LiteLLM latency > 1000ms alert to Hermes