kserve ingress update

This commit is contained in:
ychangkim
2025-07-21 15:02:45 +09:00
parent d872a6dc6f
commit 7d61034bad
4 changed files with 34 additions and 32 deletions
+5 -5
View File
@@ -18,13 +18,13 @@
### 1) InferenceService 설정
| Name | 설명 | 기본값 |
| --------------------------- | ----------------------------------------- | --------------------- |
| ----------------------- | --------------------- | ------------------- |
| `inferenceService.name` | InferenceService 이름 | `huggingface-model` |
### 2) 모델 설정
| Name | 설명 | 기본값 |
| --------------------------- | ----------------------------------------- | -------------------------------- |
| ----------------------------------- | --------------------- | ---------------------------- |
| `inferenceService.model.args[0]` | 모델 서빙 백엔드 설정 | `--backend=vllm` |
| `inferenceService.model.args[1]` | 사용할 모델 이름 | `--model_name=phi-1.5` |
| `inferenceService.model.args[2]` | 데이터 타입 설정 | `--dtype=float16` |
@@ -33,7 +33,7 @@
### 3) 리소스 설정
| Name | 설명 | 기본값 |
| --------------------------------------- | ------------------------------ | ---------- |
| ---------------------------------------------------- | ------------------ | ------ |
| `inferenceService.resources.limits.cpu` | CPU 리소스 제한 | `8` |
| `inferenceService.resources.limits.memory` | 메모리 리소스 제한 | `16Gi` |
| `inferenceService.resources.limits.nvidia.com/gpu` | GPU 리소스 제한 | `1` |
@@ -44,9 +44,9 @@
### 4) Ingress 설정
| Name | 설명 | 기본값 |
| --------------------------------------- | ------------------------------ | ------------------------------------- |
| ---------------------------------------------------- | -------------------------------------- | ------------------------------ |
| `ingress.annotations.cert-manager.io/cluster-issuer` | TLS 인증서 발급을 위한 클러스터 이슈어 | `root-ca-issuer` |
| `ingress.host.domain` | Ingress 호스트 도메인 | `kserve.example.com` |
| `ingress.hosts[0].host` | Ingress 호스트 | `phi15.default.hopt.paasup.io` |
| `ingress.tls.secretName` | TLS 인증서 시크릿 이름 | `custom-kserve-tls` |
## 3. 참고 사항
+2 -2
View File
@@ -22,7 +22,7 @@ inferenceService:
ingress:
annotations:
cert-manager.io/cluster-issuer: "root-ca-issuer"
host:
domain: "phi15.default.hopt.paasup.io"
hosts:
- host: "phi15.default.hopt.paasup.io"
tls:
secretName: custom-kserve-tls
+8 -5
View File
@@ -11,21 +11,24 @@ metadata:
spec:
ingressClassName: {{ .Values.ingress.className }}
rules:
- host: {{ if .Values.ingress.host.wildcard }}*.{{ end }}{{ .Values.ingress.host.domain }}
{{- range .Values.ingress.hosts }}
- host: {{ .host }}
http:
paths:
- backend:
service:
name: {{ .Values.ingress.backend.service.name }}
name: {{ $.Values.ingress.backend.service.name }}
port:
number: {{ .Values.ingress.backend.service.port }}
number: {{ $.Values.ingress.backend.service.port }}
path: /
pathType: Prefix
{{- end }}
{{- if .Values.ingress.tls.enabled }}
tls:
- hosts:
- {{ .Values.ingress.host.domain }}
- {{ if .Values.ingress.host.wildcard }}*.{{ end }}{{ .Values.ingress.host.domain }}
{{- range .Values.ingress.hosts }}
- {{ .host }}
{{- end }}
secretName: {{ .Values.ingress.tls.secretName }}
{{- end }}
{{- end }}
+2 -3
View File
@@ -41,9 +41,8 @@ ingress:
className: "kong"
annotations:
cert-manager.io/cluster-issuer: kubeflow-self-signing-issuer
host:
domain: "demo01-test1.hopt.paasup.io"
wildcard: false
hosts:
- host: "demo01-test1.hopt.paasup.io"
backend:
service:
name: istio-ingressgateway