SyslogSolution/syslog-harness
7
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 1 Wiki Activity
134 Commits 4 Branches 1 Tag
ct116-inference-analysis-20260912
Commit Graph
8 Commits
Author SHA1 Message Date
agent-zero 93e7605aa2 chore(ct116): capture production state (LiteLLM 1.99.1, nginx /ui //docs, router decommission, trove agent) 2026-09-11 19:34:30 +00:00
Abiba 7cad063e27 fix(routing): restore gpu-dense (.8 qwen3.6-27B-code) as syslog-auto 0.55-weight member 
The Aug 24 rename moved the syslog-auto top-weight member to .110/gpu-vision,
bypassing gpu-dense entirely (0 requests/hr while strix-moe and .110 absorbed
everything). Restore the intended 55/30/15 split: gpu-dense, strix-moe, .110.
 2026-08-28 14:53:51 +00:00
Abiba 51f0840426 feat: per-agent API keys + model pricing + nginx pass-through 2026-06-28 19:03:41 +00:00
Abiba e9aa220ee8 feat(gpu-fleet): roster-driven GPU config + Ornith-1.0-35B 2026-06-28 16:24:54 +00:00
Abiba 08680b0f9e fix: LiteLLM OIDC + Admin UI fixes - Authentik integration restored 
- Added extra_hosts for auth.sysloggh.net to LiteLLM container
- Fixed DOCS_URL=/docs (was /litellm/docs - path mismatch)
- Added Authentik self-signed cert to CA bundle
- Added nginx auth proxy for token/userinfo endpoints (SSL verify off)
- Changed OIDC token/userinfo endpoints to use nginx internal proxy
- Admin UI serving correctly on :4001/ui/ and /litellm/ui/
- Swagger API docs working at /docs and /litellm/docs
- ReDoc API docs working at /redoc and /litellm/redoc
- OIDC login flow verified working end-to-end
 2026-06-25 20:33:22 +00:00
Abiba fd3c2a575a feat: 2-layer architecture foundation + agent migration 
Router v2:
- Atomic GPU slot booking (Redis Lua — closes TOCTOU race)
- Circuit breaker with failure threshold (3 failures/120s → 60s cooldown)
- GPU health scoring with configurable weights (VRAM 40% + temp 30% + load 30%)
- X-Usage-Tokens header for LiteLLM spend tracking
- /health/unified endpoint (aggregates all layers)
- Strict explicit model passthrough (no silent fallback)
- syslog-auto → auto routing fix

Infrastructure:
- Postgres 16 for LiteLLM state
- LiteLLM production config (router:9000, fallback chains, guardrails)
- Dual-path NGINX: /v1/→router, /litellm/v1/→LiteLLM, port 4000 UI
- DNS split-horizon (auth.sysloggh.net → 192.168.68.11)
- 6 LiteLLM virtual keys for agent cutover

Deployed to CT 116, all 6 containers healthy.
 2026-06-17 22:40:33 +00:00
Abiba 85608d7c60 Dashboard + LiteLLM config updates from maintenance 2026-06-07 22:49:50 +00:00
Abiba 28fc57c5c7 May 19, 2026: Full harness update 
- Model migration: gemma-4-E4B → qwen3.5-9b-vlm
- Dashboard reorder: Usage Over Time + GPU Metrics to top
- Router counter leak fix (gpu_decr in except handler)
- VLM slot upgrade 1→2
- Automated maintenance cron job
- LiteLLM config update
 2026-05-19 15:03:47 +00:00