release 2.8.4 (#71)

helm-charts/seldon-core-v2-crds/Chart.yaml

@@ -1,9 +1,9 @@
 apiVersion: v1
-appVersion: 2.8.3
+appVersion: 2.8.4
 description: Seldon Core V2 CRDs
 keywords:
 - kubernetes
 - machine-learning
 - mlops
 name: seldon-core-v2-crds
-version: 2.8.3
+version: 2.8.4

helm-charts/seldon-core-v2-crds/templates/seldon-v2-crds.yaml

@@ -341,12 +341,17 @@ spec:
                 description: Total number of replicas targeted by this model
                 format: int32
                 type: integer
+              selector:
+                type: string
+            required:
+            - selector
             type: object
         type: object
     served: true
     storage: true
     subresources:
       scale:
+        labelSelectorPath: .status.selector
         specReplicasPath: .spec.replicas
         statusReplicasPath: .status.replicas
       status: {}
@@ -636,6 +641,14 @@ spec:
               components:
                 items:
                   properties:
+                    annotations:
+                      additionalProperties:
+                        type: string
+                      type: object
+                    labels:
+                      additionalProperties:
+                        type: string
+                      type: object
                     name:
                       type: string
                     podSpec:

helm-charts/seldon-core-v2-runtime/Chart.yaml

@@ -1,9 +1,9 @@
 apiVersion: v1
-appVersion: 2.8.3
+appVersion: 2.8.4
 description: Seldon Core V2 Runtime
 keywords:
 - kubernetes
 - machine-learning
 - mlops
 name: seldon-core-v2-runtime
-version: 2.8.3
+version: 2.8.4

helm-charts/seldon-core-v2-runtime/templates/seldon-runtime.yaml

@@ -10,22 +10,28 @@ spec:
   - name: hodometer
     disable: {{ .Values.hodometer.disable }}
     replicas: {{ .Values.hodometer.replicas }}
+    podSpec: {{ toJson .Values.hodometer.podSpec }}
   - name: seldon-scheduler
     disable: {{ .Values.scheduler.disable }}
     serviceType: {{ .Values.scheduler.serviceType }}
+    podSpec: {{ toJson .Values.scheduler.podSpec }}
   - name: seldon-envoy
     disable: {{ .Values.envoy.disable }}
     replicas: {{ .Values.envoy.replicas }}
     serviceType: {{ .Values.envoy.serviceType }}
+    podSpec: {{ toJson .Values.envoy.podSpec }}
   - name: seldon-dataflow-engine
     disable: {{ .Values.dataflow.disable }}
     replicas: {{ .Values.dataflow.replicas }}
+    podSpec: {{ toJson .Values.dataflow.podSpec }}
   - name: seldon-modelgateway
     disable: {{ .Values.modelgateway.disable }}
     replicas: {{ .Values.modelgateway.replicas }}
+    podSpec: {{ toJson .Values.modelgateway.podSpec }}
   - name: seldon-pipelinegateway
     disable: {{ .Values.pipelinegateway.disable }}
     replicas: {{ .Values.pipelinegateway.replicas }}
+    podSpec: {{ toJson .Values.pipelinegateway.podSpec }}
   config:
     agentConfig:
       rclone:

helm-charts/seldon-core-v2-servers/Chart.yaml

@@ -1,9 +1,9 @@
 apiVersion: v1
-appVersion: 2.8.3
+appVersion: 2.8.4
 description: Seldon Core V2 Servers
 keywords:
 - kubernetes
 - machine-learning
 - mlops
 name: seldon-core-v2-servers
-version: 2.8.3
+version: 2.8.4

helm-charts/seldon-core-v2-servers/templates/seldon-v2-servers.yaml

@@ -1,17 +1,27 @@
 apiVersion: mlops.seldon.io/v1alpha1
 kind: Server
 metadata:
+  annotations:
+    {{ .Values.mlserver.annotations | toYaml }}
+  labels:
+    {{ .Values.mlserver.labels | toYaml }}
   name: mlserver
   namespace: '{{ .Release.Namespace }}'
 spec:
+  podSpec: {{ toJson .Values.mlserver.podSpec }}
   replicas: {{ .Values.mlserver.replicas }}
   serverConfig: mlserver
 ---
 apiVersion: mlops.seldon.io/v1alpha1
 kind: Server
 metadata:
+  annotations:
+    {{ .Values.triton.annotations | toYaml }}
+  labels:
+    {{ .Values.triton.labels | toYaml }}
   name: triton
   namespace: '{{ .Release.Namespace }}'
 spec:
+  podSpec: {{ toJson .Values.triton.podSpec }}
   replicas: {{ .Values.triton.replicas }}
   serverConfig: triton

helm-charts/seldon-core-v2-setup/Chart.yaml

@@ -1,9 +1,9 @@
 apiVersion: v1
-appVersion: 2.8.3
+appVersion: 2.8.4
 description: Seldon Core V2 Setup
 keywords:
 - kubernetes
 - machine-learning
 - mlops
 name: seldon-core-v2-setup
-version: 2.8.3
+version: 2.8.4

helm-charts/seldon-core-v2-setup/templates/seldon-v2-components.yaml

@@ -894,7 +894,10 @@ metadata:
   namespace: '{{ .Release.Namespace }}'
 spec:
   components:
-  - name: seldon-scheduler
+  - annotations: {{- toYaml .Values.scheduler.annotations | nindent
+      8 }}
+    labels: {{- toYaml .Values.scheduler.labels | nindent 8 }}
+    name: seldon-scheduler
     podSpec:
       containers:
       - args:
@@ -903,6 +906,7 @@ spec:
         - --db-path=/mnt/scheduler/db
         - --allow-plaintxt=$(ALLOW_PLAINTXT)
         - --kafka-config-path=/mnt/kafka/kafka.json
+        - --scheduler-ready-timeout-seconds=$(SCHEDULER_READY_TIMEOUT_SECONDS)
         command:
         - /bin/scheduler
         env:
@@ -949,6 +953,8 @@ spec:
           value: '{{ .Values.security.envoy.ssl.downstream.server.caPath }}'
         - name: ENVOY_DOWNSTREAM_CLIENT_TLS_CA_LOCATION
           value: '{{ .Values.security.envoy.ssl.downstream.server.clientCaPath }}'
+        - name: SCHEDULER_READY_TIMEOUT_SECONDS
+          value: '{{ .Values.scheduler.schedulerReadyTimeoutSeconds }}'
         - name: ALLOW_PLAINTXT
           value: "true"
         - name: POD_NAMESPACE
@@ -1009,7 +1015,10 @@ spec:
         resources:
           requests:
             storage: '{{ .Values.scheduler.resources.storage }}'
-  - name: seldon-pipelinegateway
+  - annotations: {{- toYaml .Values.pipelinegateway.annotations | nindent
+      8 }}
+    labels: {{- toYaml .Values.pipelinegateway.labels | nindent 8 }}
+    name: seldon-pipelinegateway
     podSpec:
       containers:
       - args:
@@ -1145,7 +1154,10 @@ spec:
           name: seldon-tracing
         name: tracing-config-volume
     replicas: 1
-  - name: seldon-modelgateway
+  - annotations: {{- toYaml .Values.modelgateway.annotations | nindent
+      8 }}
+    labels: {{- toYaml .Values.modelgateway.labels | nindent 8 }}
+    name: seldon-modelgateway
     podSpec:
       containers:
       - args:
@@ -1263,7 +1275,10 @@ spec:
           name: seldon-tracing
         name: tracing-config-volume
     replicas: 1
-  - name: hodometer
+  - annotations: {{- toYaml .Values.hodometer.annotations | nindent
+      8 }}
+    labels: {{- toYaml .Values.hodometer.labels | nindent 8 }}
+    name: hodometer
     podSpec:
       containers:
       - env:
@@ -1319,7 +1334,9 @@ spec:
       serviceAccountName: hodometer
       terminationGracePeriodSeconds: 5
     replicas: 1
-  - name: seldon-envoy
+  - annotations: {{- toYaml .Values.envoy.annotations | nindent 8 }}
+    labels: {{- toYaml .Values.envoy.labels | nindent 8 }}
+    name: seldon-envoy
     podSpec:
       containers:
       - env:
@@ -1374,7 +1391,10 @@ spec:
         8 }}
       terminationGracePeriodSeconds: 5
     replicas: 1
-  - name: seldon-dataflow-engine
+  - annotations: {{- toYaml .Values.dataflow.annotations | nindent 8
+      }}
+    labels: {{- toYaml .Values.dataflow.labels | nindent 8 }}
+    name: seldon-dataflow-engine
     podSpec:
       containers:
       - env:
@@ -1539,6 +1559,24 @@ spec:
         value: '{{ .Values.serverConfig.agent.modelInactiveSecondsThreshold }}'
       - name: SELDON_SCALING_STATS_PERIOD_SECONDS
         value: '{{ .Values.serverConfig.agent.scalingStatsPeriodSeconds }}'
+      - name: SELDON_MAX_TIME_READY_SUB_SERVICE_AFTER_START_SECONDS
+        value: '{{ .Values.serverConfig.agent.maxElapsedTimeReadySubServiceAfterStartSeconds
+          }}'
+      - name: SELDON_MAX_ELAPSED_TIME_READY_SUB_SERVICE_BEFORE_START_MINUTES
+        value: '{{ .Values.serverConfig.agent.maxElapsedTimeReadySubServiceBeforeStartMinutes
+          }}'
+      - name: SELDON_PERIOD_READY_SUB_SERVICE_SECONDS
+        value: '{{ .Values.serverConfig.agent.periodReadySubServiceSeconds }}'
+      - name: SELDON_MAX_LOAD_ELAPSED_TIME_MINUTES
+        value: '{{ .Values.serverConfig.agent.maxLoadElapsedTimeMinutes }}'
+      - name: SELDON_MAX_UNLOAD_ELAPSED_TIME_MINUTES
+        value: '{{ .Values.serverConfig.agent.maxUnloadElapsedTimeMinutes }}'
+      - name: SELDON_MAX_LOAD_RETRY_COUNT
+        value: '{{ .Values.serverConfig.agent.maxLoadRetryCount }}'
+      - name: SELDON_MAX_UNLOAD_RETRY_COUNT
+        value: '{{ .Values.serverConfig.agent.maxUnloadRetryCount }}'
+      - name: SELDON_UNLOAD_GRACE_PERIOD_SECONDS
+        value: '{{ .Values.serverConfig.agent.unloadGracePeriodSeconds }}'
       - name: SELDON_OVERCOMMIT_PERCENTAGE
         value: '{{ .Values.serverConfig.agent.overcommitPercentage }}'
       - name: CONTROL_PLANE_SECURITY_PROTOCOL
@@ -1787,6 +1825,24 @@ spec:
         value: '{{ .Values.serverConfig.agent.modelInactiveSecondsThreshold }}'
       - name: SELDON_SCALING_STATS_PERIOD_SECONDS
         value: '{{ .Values.serverConfig.agent.scalingStatsPeriodSeconds }}'
+      - name: SELDON_MAX_TIME_READY_SUB_SERVICE_AFTER_START_SECONDS
+        value: '{{ .Values.serverConfig.agent.maxElapsedTimeReadySubServiceAfterStartSeconds
+          }}'
+      - name: SELDON_MAX_ELAPSED_TIME_READY_SUB_SERVICE_BEFORE_START_MINUTES
+        value: '{{ .Values.serverConfig.agent.maxElapsedTimeReadySubServiceBeforeStartMinutes
+          }}'
+      - name: SELDON_PERIOD_READY_SUB_SERVICE_SECONDS
+        value: '{{ .Values.serverConfig.agent.periodReadySubServiceSeconds }}'
+      - name: SELDON_MAX_LOAD_ELAPSED_TIME_MINUTES
+        value: '{{ .Values.serverConfig.agent.maxLoadElapsedTimeMinutes }}'
+      - name: SELDON_MAX_UNLOAD_ELAPSED_TIME_MINUTES
+        value: '{{ .Values.serverConfig.agent.maxUnloadElapsedTimeMinutes }}'
+      - name: SELDON_MAX_LOAD_RETRY_COUNT
+        value: '{{ .Values.serverConfig.agent.maxLoadRetryCount }}'
+      - name: SELDON_MAX_UNLOAD_RETRY_COUNT
+        value: '{{ .Values.serverConfig.agent.maxUnloadRetryCount }}'
+      - name: SELDON_UNLOAD_GRACE_PERIOD_SECONDS
+        value: '{{ .Values.serverConfig.agent.unloadGracePeriodSeconds }}'
       - name: SELDON_OVERCOMMIT_PERCENTAGE
         value: '{{ .Values.serverConfig.agent.overcommitPercentage }}'
       - name: CONTROL_PLANE_SECURITY_PROTOCOL

helm-charts/seldon-core-v2-setup/values.yaml

@@ -84,7 +84,7 @@ hodometer:
     pullPolicy: IfNotPresent
     registry: docker.io
     repository: seldonio/seldon-hodometer
-    tag: 2.8.3
+    tag: 2.8.4
   logLevel: info
   metricsLevel: feature
   extraPublishUrls: ""
@@ -121,7 +121,7 @@ modelgateway:
     pullPolicy: IfNotPresent
     registry: docker.io
     repository: seldonio/seldon-modelgateway
-    tag: 2.8.3
+    tag: 2.8.4
   workers: 8
   resources:
     cpu: 100m
@@ -137,7 +137,7 @@ pipelinegateway:
     pullPolicy: IfNotPresent
     registry: docker.io
     repository: seldonio/seldon-pipelinegateway
-    tag: 2.8.3
+    tag: 2.8.4
   resources:
     cpu: 100m
     memory: 1G
@@ -152,7 +152,7 @@ dataflow:
     pullPolicy: IfNotPresent
     registry: docker.io
     repository: seldonio/seldon-dataflow-engine
-    tag: 2.8.3
+    tag: 2.8.4
   cores: 4
   resources:
     cpu: 100m
@@ -169,7 +169,7 @@ controller:
     pullPolicy: IfNotPresent
     registry: docker.io
     repository: seldonio/seldonv2-controller
-    tag: 2.8.3
+    tag: 2.8.4
   resources:
     cpu: 10m
     memory: 64Mi
@@ -180,11 +180,15 @@ controller:
     runAsNonRoot: true
 
 envoy:
+  annotations:
+    "prometheus.io/path": "/stats/prometheus"
+    "prometheus.io/port": "9003"
+    "prometheus.io/scrape": "true"
   image:
     pullPolicy: IfNotPresent
     registry: docker.io
     repository: seldonio/seldon-envoy
-    tag: 2.8.3
+    tag: 2.8.4
   resources:
     cpu: 100m
     memory: 128Mi
@@ -201,7 +205,7 @@ scheduler:
     pullPolicy: IfNotPresent
     registry: docker.io
     repository: seldonio/seldon-scheduler
-    tag: 2.8.3
+    tag: 2.8.4
   resources:
     cpu: 100m
     memory: 1Gi
@@ -213,6 +217,7 @@ scheduler:
     runAsUser: 1000
     runAsGroup: 1000
     runAsNonRoot: true
+  schedulerReadyTimeoutSeconds: 600
 
 serverConfig:
   terminationGracePeriodSeconds: 120
@@ -228,7 +233,7 @@ serverConfig:
       pullPolicy: IfNotPresent
       registry: docker.io
       repository: seldonio/seldon-rclone
-      tag: 2.8.3
+      tag: 2.8.4
     resources:
       cpu: 50m
       memory: 128Mi
@@ -238,11 +243,19 @@ serverConfig:
       pullPolicy: IfNotPresent
       registry: docker.io
       repository: seldonio/seldon-agent
-      tag: 2.8.3
+      tag: 2.8.4
     overcommitPercentage: "10"
     scalingStatsPeriodSeconds: "20"
     modelInferenceLagThreshold: "30"
     modelInactiveSecondsThreshold: "600"
+    maxElapsedTimeReadySubServiceAfterStartSeconds: "30"
+    maxElapsedTimeReadySubServiceBeforeStartMinutes: "15"
+    periodReadySubServiceSeconds: "60"
+    maxLoadElapsedTimeMinutes: "120"
+    maxUnloadElapsedTimeMinutes: "15"
+    maxLoadRetryCount: "5"
+    maxUnloadRetryCount: "1"
+    unloadGracePeriodSeconds: "2"
     resources:
       cpu: 200m
       memory: 1Gi