fix(chart): .NET 后端 probe 改极致保守值 + submodule 指针

_backend.tpl probe 原值(liveness delay=15 period=30,无 timeout/fail)
走 K8s 默认 timeout=1s fail=3,约 45s 即杀,.NET 冷启动连库常超时。

改为极致保守:
- liveness: delay=120 timeout=15 period=30 fail=12 → 480s 才杀
- readiness: delay=60 timeout=10 period=30 fail=8 → 300s 才标 NotReady

影响 sso/agent-0/edge-0/ultra/gateway 共 5 个 .NET 服务。
helm template 验证:5 个 livenessProbe 全带上新值,无渲染错误。

同步更新 mixingx.edge submodule 指针(opcua Dockerfile ENTRYPOINT 修正)。
This commit is contained in:
ximenchuifeng committed 2026-06-26 08:28:25 +08:00
1 parent 8ffc6ee825
commit 92027fc81c
1 file changed
+7 -3
+7 -3
View File
@@ -69,15 +69,19 @@ spec:
path: /health
port: http
scheme: HTTP
initialDelaySeconds: 15
initialDelaySeconds: 120
periodSeconds: 30
timeoutSeconds: 15
failureThreshold: 12 # 120 + 30×12 = 480s 才杀,容忍 .NET 冷启动连库
readinessProbe:
httpGet:
path: /alive
port: http
scheme: HTTP
initialDelaySeconds: 5
periodSeconds: 10
initialDelaySeconds: 60
periodSeconds: 30
timeoutSeconds: 10
failureThreshold: 8 # 60 + 30×8 = 300s 才标 NotReady
---
# Service: {{ .name }}