docs(logbook): record ollama policy order deploy [skip ci]
This commit is contained in:
@@ -14,6 +14,31 @@
|
||||
- `OLLAMA_FALLBACK_URL=http://192.168.0.110:11437`
|
||||
- 新增 `test_ollama_prod_manifest_order.py`,把 production manifest 的 Ollama policy order 鎖成測試,避免再出現 `ollama_gcp_a` label 指到 GCP-B/111。
|
||||
|
||||
**deploy / live verification**:
|
||||
|
||||
```text
|
||||
Gitea:
|
||||
code-review #2114 -> success
|
||||
cd #2113 -> success
|
||||
production:
|
||||
image -> 192.168.0.110:5000/awoooi/api:b9fc8748a5d3e03ca585779a39c8b07af22334de
|
||||
OLLAMA_URL=http://192.168.0.110:11435
|
||||
OLLAMA_SECONDARY_URL=http://192.168.0.110:11436
|
||||
OLLAMA_FALLBACK_URL=http://192.168.0.110:11437
|
||||
ai-route-status:
|
||||
policy_order -> ollama_gcp_a(11435) -> ollama_gcp_b(11436) -> ollama_local(11437) -> gemini
|
||||
selected_provider -> ollama_gcp_b
|
||||
selected_url -> http://192.168.0.110:11436
|
||||
health:
|
||||
status -> degraded
|
||||
ollama -> primary unavailable; fallback active: ollama_gcp_b
|
||||
ollama_gcp_a -> down, recent endpoint failure cooldown
|
||||
ollama_gcp_b/local -> up
|
||||
log sample:
|
||||
5x health + 5x ai-route-status calls:
|
||||
/var/log/nginx/ollama-gcp-a-error.log delta -> +2 lines, not +10
|
||||
```
|
||||
|
||||
**注意**:
|
||||
|
||||
- 這不是否認 GCP-B 可用;正確流程是 policy 先顯示 GCP-A red,再由 failover manager 選 GCP-B 或 111。
|
||||
|
||||
Reference in New Issue
Block a user