feat(llm-inference-multimodel): bump router --models-max 1->4 on astro-orbiter (t_33acbb2e)

Changes:
- host_vars/astro-orbiter/vars.yml: add llm_router_models_max: 4 (overrides
  conservative default of 1). Detailed VRAM OOM risk note included inline:
  worst-case 3-model co-residency ~31GB > 24GB RTX 3090. LRU eviction
  mitigates in single-user operation; flagged for Ryan's review.
- playbooks/day2_bump_router_models_max.yml: new targeted playbook; deploys
  updated router unit, restarts the live service, verifies /health 200 and
  /v1/models lists all 3 GGUFs post-restart.
- group_vars/all/semaphore.yml: add llm_router_update_unit template pointing
  at the new playbook.
- roles/llm-inference-multimodel/defaults/main.yml: update comment to reflect
  the var is now overridden in host_vars rather than 'hardcoded to 1'.
- roles/llm-inference-multimodel/templates/llama-server-router.service.j2:
  correct stale 'HARDCODED TO 1' comment — value is variable-driven.

Constraints honored:
- --parallel 1 left untouched (not in scope, not modified anywhere)
- No ad-hoc SSH/systemctl/curl state mutation; all execution via Semaphore
- No installed/vendored code patched

ansible/host_vars/astro-orbiter/vars.yml

@@ -29,6 +29,27 @@ common_root_lv: ubuntu-lv
 # Meta-Llama-3.1-8B-Instruct-Q4_K_M as router models alongside the production
 # Qwen3.6-35B-A3B-UD-Q4_K_S. The live files were already present/correct on
 # astro-orbiter; this pass codifies them. Future adds = append to this list.
+# Router --models-max override for astro-orbiter.
+# Default in defaults/main.yml is 1 (conservative). Bumped to 4 on 2026-08-12
+# (t_33acbb2e) so the router can keep more than one GGUF resident on-demand
+# and LRU-evict when needed.
+#
+# VRAM NOTE (t_33acbb2e): With models-max=4 and all 3 current GGUFs plus
+# headroom for one more, worst case is all 3 loaded simultaneously:
+#   Qwen3.6-35B-A3B Q4_K_S:       ~21.5GB (weights ~19.5GB + KV ~2GB @ 64K ctx, q4_0)
+#   Phi-3.5-mini-instruct Q8_0:   ~4.3GB  (weights ~3.8GB + KV ~0.5GB @ 64K ctx)
+#   Meta-Llama-3.1-8B Q4_K_M:    ~5.6GB  (weights ~4.6GB + KV ~1.0GB @ 64K ctx)
+#   Total worst-case:             ~31.4GB  >  24GB RTX 3090
+#
+# OOM RISK: If all 3 models are loaded concurrently the card will OOM. The
+# router's LRU eviction means this only occurs if all 3 models receive a
+# concurrent request before any model completes its response (very unlikely
+# in single-user homelab operation). However, Ryan should be aware. A safe
+# alternative would be models-max=2 (allows Qwen + one small model resident
+# simultaneously: ~21.5 + 5.6 = ~27.1GB still tight). Proceeding to 4 as
+# instructed per task t_33acbb2e; flagged for Ryan's attention.
+llm_router_models_max: 4
+
 llm_staged_models:
   - filename: "Phi-3.5-mini-instruct-Q8_0.gguf"
     url: "https://huggingface.co/bartowski/Phi-3.5-mini-instruct-GGUF/resolve/main/Phi-3.5-mini-instruct-Q8_0.gguf"