Bug fixes for replicas, node selectors, tolerations and secrets

.gitignore

@@ -0,0 +1,3 @@
+# IDE files
+.idea/
+.vscode/

charts/clearml-serving/Chart.yaml

@@ -2,7 +2,7 @@ apiVersion: v2
 name: clearml-serving
 description: ClearML Serving Helm Chart
 type: application
-version: "1.6.2"
+version: "1.6.3"
 appVersion: "1.3.0"
 kubeVersion: ">= 1.21.0-0"
 home: https://clear.ml
@@ -34,4 +34,4 @@ dependencies:
 annotations:
   artifacthub.io/changes: |
     - kind: fixed
-      description: "Bitnami registries workaround"
+      description: "Fixes for issue #375 - nodeSelector/toleration for inference app, replicas, etc."

charts/clearml-serving/README.md

@@ -1,6 +1,6 @@
 # ClearML Kubernetes Serving
 
-![Version: 1.6.2](https://img.shields.io/badge/Version-1.6.2-informational?style=flat-square) ![Type: application](https://img.shields.io/badge/Type-application-informational?style=flat-square) ![AppVersion: 1.3.0](https://img.shields.io/badge/AppVersion-1.3.0-informational?style=flat-square)
+![Version: 1.6.3](https://img.shields.io/badge/Version-1.6.3-informational?style=flat-square) ![Type: application](https://img.shields.io/badge/Type-application-informational?style=flat-square) ![AppVersion: 1.3.0](https://img.shields.io/badge/AppVersion-1.3.0-informational?style=flat-square)
 
 ClearML Serving Helm Chart
 
@@ -61,8 +61,9 @@ Kubernetes: `>= 1.21.0-0`
 
 | Key | Type | Default | Description |
 |-----|------|---------|-------------|
-| clearml | object | `{"apiAccessKey":"ClearML API Access Key","apiHost":"http://clearml-server-apiserver:8008","apiSecretKey":"ClearML API Secret Key","defaultBaseServeUrl":"http://127.0.0.1:8080/serve","filesHost":"http://clearml-server-fileserver:8081","kafkaServeUrl":"","servingTaskId":"ClearML Serving Task ID","webHost":"http://clearml-server-webserver:80"}` | ClearMl generic configurations |
-| clearml_serving_inference | object | `{"additionalConfigs":{},"affinity":{},"autoscaling":{"enabled":false,"maxReplicas":11,"minReplicas":1,"targetCPU":50,"targetMemory":50},"existingAdditionalConfigsConfigMap":"","existingAdditionalConfigsSecret":"","extraEnvironment":[],"extraPythonPackages":[],"image":{"repository":"allegroai/clearml-serving-inference","tag":"1.3.0"},"ingress":{"annotations":{},"enabled":false,"hostName":"serving.clearml.127-0-0-1.nip.io","ingressClassName":"","path":"/","tlsSecretName":""},"nodeSelector":{},"pollFrequencyMinutes":"1.0","resources":{},"tolerations":[]}` | ClearML serving inference configurations |
+| clearml | object | `{"apiAccessKey":"ClearML API Access Key","apiHost":"http://clearml-server-apiserver:8008","apiKeyExistingSecret":"","apiSecretKey":"ClearML API Secret Key","defaultBaseServeUrl":"http://127.0.0.1:8080/serve","filesHost":"http://clearml-server-fileserver:8081","kafkaServeUrl":"","servingTaskId":"ClearML Serving Task ID","webHost":"http://clearml-server-webserver:80"}` | ClearMl generic configurations |
+| clearml.apiKeyExistingSecret | string | `""` | Pass ClearML API secrets using an existing secret. The secret must contain fields: api-access-key, api-secret-key |
+| clearml_serving_inference | object | `{"additionalConfigs":{},"affinity":{},"autoscaling":{"enabled":false,"maxReplicas":11,"minReplicas":1,"targetCPU":50,"targetMemory":50},"existingAdditionalConfigsConfigMap":"","existingAdditionalConfigsSecret":"","extraEnvironment":[],"extraPythonPackages":[],"image":{"repository":"allegroai/clearml-serving-inference","tag":"1.3.0"},"ingress":{"annotations":{},"enabled":false,"hostName":"serving.clearml.127-0-0-1.nip.io","ingressClassName":"","path":"/","tlsSecretName":""},"nodeSelector":{},"pollFrequencyMinutes":"1.0","replicas":1,"resources":{},"tolerations":[]}` | ClearML serving inference configurations |
 | clearml_serving_inference.additionalConfigs | object | `{}` | files declared in this parameter will be mounted on internal folder /opt/clearml/config and read by pod (examples in values.yaml) if not overridden by existingAdditionalConfigsSecret |
 | clearml_serving_inference.affinity | object | `{}` | Affinity configuration |
 | clearml_serving_inference.autoscaling | object | `{"enabled":false,"maxReplicas":11,"minReplicas":1,"targetCPU":50,"targetMemory":50}` | Autoscaling configuration |
@@ -80,9 +81,10 @@ Kubernetes: `>= 1.21.0-0`
 | clearml_serving_inference.ingress.tlsSecretName | string | `""` | Reference to secret containing TLS certificate. If set, it enables HTTPS on ingress rule. |
 | clearml_serving_inference.nodeSelector | object | `{}` | Node Selector configuration |
 | clearml_serving_inference.pollFrequencyMinutes | string | `"1.0"` | Default polling frequency for finding new models |
+| clearml_serving_inference.replicas | int | `1` | Number of replicas - only effective when autoscaling is disabled |
 | clearml_serving_inference.resources | object | `{}` | Pod resources definition |
 | clearml_serving_inference.tolerations | list | `[]` | Tolerations configuration |
-| clearml_serving_statistics | object | `{"additionalConfigs":{},"affinity":{},"enabled":true,"existingAdditionalConfigsConfigMap":"","existingAdditionalConfigsSecret":"","extraEnvironment":[],"extraPythonPackages":[],"image":{"repository":"allegroai/clearml-serving-statistics","tag":"1.3.0"},"nodeSelector":{},"pollFrequencyMinutes":"1.0","resources":{},"tolerations":[]}` | ClearML serving statistics configurations |
+| clearml_serving_statistics | object | `{"additionalConfigs":{},"affinity":{},"enabled":true,"existingAdditionalConfigsConfigMap":"","existingAdditionalConfigsSecret":"","extraEnvironment":[],"extraPythonPackages":[],"image":{"repository":"allegroai/clearml-serving-statistics","tag":"1.3.0"},"nodeSelector":{},"pollFrequencyMinutes":"1.0","replicas":1,"resources":{},"tolerations":[]}` | ClearML serving statistics configurations |
 | clearml_serving_statistics.additionalConfigs | object | `{}` | files declared in this parameter will be mounted on internal folder /opt/clearml/config and read by pod (examples in values.yaml) if not overridden by existingAdditionalConfigsSecret |
 | clearml_serving_statistics.affinity | object | `{}` | Affinity configuration |
 | clearml_serving_statistics.enabled | bool | `true` | Enable ClearML Serving Statistics |
@@ -92,9 +94,10 @@ Kubernetes: `>= 1.21.0-0`
 | clearml_serving_statistics.image | object | `{"repository":"allegroai/clearml-serving-statistics","tag":"1.3.0"}` | Container Image |
 | clearml_serving_statistics.nodeSelector | object | `{}` | Node Selector configuration |
 | clearml_serving_statistics.pollFrequencyMinutes | string | `"1.0"` | Default polling frequency for finding new models |
+| clearml_serving_statistics.replicas | int | `1` | Number of replicas |
 | clearml_serving_statistics.resources | object | `{}` | Pod resources definition |
 | clearml_serving_statistics.tolerations | list | `[]` | Tolerations configuration |
-| clearml_serving_triton | object | `{"additionalConfigs":{},"affinity":{},"autoscaling":{"enabled":false,"maxReplicas":11,"minReplicas":1,"targetCPU":50,"targetMemory":50},"enabled":true,"existingAdditionalConfigsConfigMap":"","existingAdditionalConfigsSecret":"","extraEnvironment":[],"extraPythonPackages":[],"image":{"repository":"allegroai/clearml-serving-triton","tag":"1.3.0"},"ingress":{"annotations":{},"enabled":false,"hostName":"serving-grpc.clearml.127-0-0-1.nip.io","ingressClassName":"","path":"/","tlsSecretName":""},"metricFrequencyMinutes":"1.0","nodeSelector":{},"pollFrequencyMinutes":"1.0","resources":{},"tolerations":[]}` | ClearML serving Triton configurations |
+| clearml_serving_triton | object | `{"additionalConfigs":{},"affinity":{},"autoscaling":{"enabled":false,"maxReplicas":11,"minReplicas":1,"targetCPU":50,"targetMemory":50},"enabled":true,"existingAdditionalConfigsConfigMap":"","existingAdditionalConfigsSecret":"","extraEnvironment":[],"extraPythonPackages":[],"image":{"repository":"allegroai/clearml-serving-triton","tag":"1.3.0"},"ingress":{"annotations":{},"enabled":false,"hostName":"serving-grpc.clearml.127-0-0-1.nip.io","ingressClassName":"","path":"/","tlsSecretName":""},"metricFrequencyMinutes":"1.0","nodeSelector":{},"pollFrequencyMinutes":"1.0","replicas":1,"resources":{},"tolerations":[]}` | ClearML serving Triton configurations |
 | clearml_serving_triton.additionalConfigs | object | `{}` | files declared in this parameter will be mounted on internal folder /opt/clearml/config and read by pod (examples in values.yaml) if not overridden by existingAdditionalConfigsSecret |
 | clearml_serving_triton.affinity | object | `{}` | Affinity configuration |
 | clearml_serving_triton.autoscaling | object | `{"enabled":false,"maxReplicas":11,"minReplicas":1,"targetCPU":50,"targetMemory":50}` | Autoscaling configuration |
@@ -114,6 +117,7 @@ Kubernetes: `>= 1.21.0-0`
 | clearml_serving_triton.metricFrequencyMinutes | string | `"1.0"` | Triton metric reporting update frequency in minutes |
 | clearml_serving_triton.nodeSelector | object | `{}` | Node Selector configuration |
 | clearml_serving_triton.pollFrequencyMinutes | string | `"1.0"` | Triton model update frequency in minutes |
+| clearml_serving_triton.replicas | int | `1` | Number of replicas - only effective when autoscaling is disabled |
 | clearml_serving_triton.resources | object | `{}` | Pod resources definition |
 | clearml_serving_triton.tolerations | list | `[]` | Tolerations configuration |
 | grafana | object | `{"adminPassword":"clearml","adminUser":"admin","datasources":{"datasources.yaml":{"apiVersion":1,"datasources":[{"access":"proxy","isDefault":true,"name":"Prometheus","type":"prometheus","url":"http://{{ .Release.Name }}-prometheus-server"}]}},"enabled":true}` | Configuration from https://github.com/grafana/helm-charts/blob/main/charts/grafana/values.yaml |

charts/clearml-serving/templates/clearml-serving-inference-deployment.yaml

@@ -6,7 +6,9 @@ metadata:
     clearml.serving.service: {{ include "clearmlServing.fullname" . }}-inference
   name: {{ include "clearmlServing.fullname" . }}-inference
 spec:
-  replicas: 1
+  {{- if not .Values.clearml_serving_inference.autoscaling.enabled }}
+  replicas: {{ .Values.clearml_serving_inference.replicas }}
+  {{- end}}
   selector:
     matchLabels:
       clearml.serving.service: {{ include "clearmlServing.fullname" . }}-inference
@@ -44,12 +46,34 @@ spec:
       affinity:
         {{- toYaml . | nindent 8 }}
       {{- end }}
+      {{- with .Values.clearml_serving_inference.nodeSelector }}
+      nodeSelector:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
+      {{- with .Values.clearml_serving_inference.tolerations }}
+      tolerations:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
       containers:
         - env:
             - name: CLEARML_API_ACCESS_KEY
+              {{- if .Values.clearml.apiKeyExistingSecret }}
+              valueFrom:
+                secretKeyRef:
+                  name: {{ .Values.clearml.apiKeyExistingSecret }}
+                  key: api-access-key
+              {{- else }}
               value: "{{ .Values.clearml.apiAccessKey }}"
+              {{- end}}
             - name: CLEARML_API_SECRET_KEY
+              {{- if .Values.clearml.apiKeyExistingSecret }}
+              valueFrom:
+                secretKeyRef:
+                  name: {{ .Values.clearml.apiKeyExistingSecret }}
+                  key: api-secret-key
+              {{- else }}
               value: "{{ .Values.clearml.apiSecretKey }}"
+              {{- end}}
             - name: CLEARML_API_HOST
               value: "{{ .Values.clearml.apiHost }}"
             - name: CLEARML_FILES_HOST
@@ -98,14 +122,6 @@ spec:
             - name: additional-config
               mountPath: /opt/clearml/config
           {{- end }}
-          {{- with .Values.clearml_serving_inference.nodeSelector }}
-          nodeSelector:
-            {{- toYaml . | nindent 12 }}
-          {{- end }}
-          {{- with .Values.clearml_serving_inference.tolerations }}
-          tolerations:
-            {{- toYaml . | nindent 12 }}
-          {{- end }}
           resources:
             {{- toYaml .Values.clearml_serving_inference.resources | nindent 12 }}
       restartPolicy: Always

charts/clearml-serving/templates/clearml-serving-statistics-deployment.yaml

@@ -7,7 +7,7 @@ metadata:
     clearml.serving.service: {{ include "clearmlServing.fullname" . }}-statistics
   name: {{ include "clearmlServing.fullname" . }}-statistics
 spec:
-  replicas: 1
+  replicas: {{ .Values.clearml_serving_statistics.replicas }}
   selector:
     matchLabels:
       clearml.serving.service: {{ include "clearmlServing.fullname" . }}-statistics

charts/clearml-serving/templates/clearml-serving-triton-deployment.yaml

@@ -7,7 +7,9 @@ metadata:
     clearml.serving.service: {{ include "clearmlServing.fullname" . }}-triton
   name: {{ include "clearmlServing.fullname" . }}-triton
 spec:
-  replicas: 1
+  {{- if not .Values.clearml_serving_triton.autoscaling.enabled }}
+  replicas: {{ .Values.clearml_serving_triton.replicas }}
+  {{- end }}
   selector:
     matchLabels:
       clearml.serving.service: {{ include "clearmlServing.fullname" . }}-triton

charts/clearml-serving/values.yaml

@@ -17,6 +17,8 @@ imageCredentials:
 clearml:
   apiAccessKey: "ClearML API Access Key"
   apiSecretKey: "ClearML API Secret Key"
+  # -- Pass ClearML API secrets using an existing secret. The secret must contain fields: api-access-key, api-secret-key
+  apiKeyExistingSecret: ""
   apiHost: http://clearml-server-apiserver:8008
   filesHost: http://clearml-server-fileserver:8081
   webHost: http://clearml-server-webserver:80
@@ -26,6 +28,8 @@ clearml:
 
 # -- ClearML serving statistics configurations
 clearml_serving_statistics:
+  # -- Number of replicas
+  replicas: 1
   # -- Enable ClearML Serving Statistics
   enabled: true
   # -- Container Image
@@ -86,6 +90,8 @@ clearml_serving_inference:
   #     <filecontent>
   # -- Default polling frequency for finding new models
   pollFrequencyMinutes: "1.0"
+  # -- Number of replicas - only effective when autoscaling is disabled
+  replicas: 1
   # -- Autoscaling configuration
   autoscaling:
     enabled: false
@@ -145,6 +151,8 @@ clearml_serving_triton:
   pollFrequencyMinutes: "1.0"
   # -- Triton metric reporting update frequency in minutes
   metricFrequencyMinutes: "1.0"
+  # -- Number of replicas - only effective when autoscaling is disabled
+  replicas: 1
   # -- Autoscaling configuration
   autoscaling:
     enabled: false