Advanced Helm Chart Configuration¶
Seldon Core Operator Chart Configuration¶
This page provides a detailed overview of the installation parameters available for the Seldon Core installation when using Helm 3.x. The high level workflows to install Seldon Core can be found in the Installation Page.
Below you can find the values.yaml file of the seldon-core-operator Helm chart, which contains basically all the values that you can configure in your installation by using the set flag in the format –set value.path=YOUR_VALUE.
The file has been written to be self documented, and has information on all the core parameters. Further information is referenced in the file to specific documentation pages.
1# # Seldon Core Operator
2# Below are the default values when installing Seldon Core
3
4# Defaults to .Release.Namespace
5namespaceOverride: ""
6
7# ## Ingress Options
8# You are able to choose between Istio and Ambassador
9
10# If you have ambassador installed you can just use the enabled flag
11ambassador:
12 enabled: false
13 version: v2
14 singleNamespace: false
15# When activating Istio, respecive virtual services will be created
16# You must make sure you create the seldon-gateway as well
17istio:
18 enabled: false
19 gateway: istio-system/seldon-gateway
20 tlsMode: ""
21# If you have KEDA installed you can use it for autoscaling
22keda:
23 enabled: false
24# ## Install with Cert Manager
25# See installation page in documentation for more information
26certManager:
27 enabled: false
28
29# ## Install with limited namespace visibility
30# If you want to ensure seldon-core-controller can only have visibility
31# to specify namespaces you can set the controllerId
32controllerId: ""
33
34# Whether operator should create the webhooks and configmap on startup (false means created from chart)
35managerCreateResources: false
36
37# Default user id to add to all Pod Security Context as the default
38# Use this to ensure all container run as non-root by default
39# For openshift leave blank as usually this will be injected automatically on an openshift cluster
40# to all pods.
41defaultUserID: "8888"
42
43# runtime user to use for manager
44managerUserID: 8888
45
46# ## Service Orchestrator (Executor)
47# The executor is the default service orchestrator which has superceded the "Java Engine"
48executor:
49 port: 8000
50 metricsPortName: metrics
51 image:
52 pullPolicy: IfNotPresent
53 registry: docker.io
54 repository: seldonio/seldon-core-executor
55 tag: 1.19.0-dev
56 resources:
57 cpuLimit: 500m
58 cpuRequest: 500m
59 memoryLimit: 512Mi
60 memoryRequest: 512Mi
61 prometheus:
62 path: /prometheus
63 serviceAccount:
64 name: default
65 user: 8888
66# If you want to make available your own request logger for ELK integration you can set this
67# For more information see the Production Integration for Payload Request Logging with ELK in the docs
68 requestLogger:
69 defaultEndpoint: 'http://default-broker'
70 workQueueSize: 10000
71 writeTimeoutMs: 2000
72 # Whether to run full protocol API based health checks on models in deployment graph. False will just do TCP connects
73 fullHealthChecks: false
74
75
76# ## Seldon Core Controller Manager Options
77image:
78 pullPolicy: IfNotPresent
79 registry: docker.io
80 repository: seldonio/seldon-core-operator
81 tag: 1.19.0-dev
82manager:
83 cpuLimit: 500m
84 cpuRequest: 100m
85 memoryLimit: 300Mi
86 memoryRequest: 200Mi
87 logLevel: INFO
88 leaderElectionID: a33bd623.machinelearning.seldon.io
89 leaderElectionResourceLock:
90 leaderElectionLeaseDurationSecs: 15
91 leaderElectionRenewDeadlineSecs: 10
92 leaderElectionRetryPeriodSecs: 2
93 annotations: {}
94 containerSecurityContext: {}
95 deploymentNameAsPrefix: false
96 priorityClassName:
97rbac:
98 configmap:
99 create: true
100 create: true
101serviceAccount:
102 create: true
103 name: seldon-manager
104singleNamespace: false
105storageInitializer:
106 cpuLimit: "1"
107 cpuRequest: 100m
108 image: seldonio/rclone-storage-initializer:1.19.0-dev
109 memoryLimit: 1Gi
110 memoryRequest: 100Mi
111usageMetrics:
112 enabled: false
113# In scenarios like EKS with non-standard CNI plugin like calico, the control plane cannot reach the webhook
114# hence it is needed to set hostNetwork: true
115hostNetwork: false
116webhook:
117 port: 4443 # If 'hostNetwork: true' you might need to change this port if it is already used by the node
118metrics: # these are the metrics exposed by the controller pod
119 port: 8080 # If 'hostNetwork: true' you might need to change this port if it is already used by the node
120
121# ## Predictive Unit Values
122predictiveUnit:
123 httpPort: 9000
124 grpcPort: 9500
125 metricsPortName: metrics
126 # If you would like to add extra environment variables to the init container to make available
127 # secrets such as cloud credentials, you can provide a default secret name that will be loaded
128 # to all the containers. You can then override this using the envSecretRefName in SeldonDeployments
129 defaultEnvSecretRefName: ""
130predictor_servers:
131 MLFLOW_SERVER:
132 protocols:
133 seldon:
134 defaultImageVersion: "1.19.0-dev"
135 image: seldonio/mlflowserver
136 v2:
137 defaultImageVersion: "1.3.5-mlflow"
138 image: seldonio/mlserver
139 SKLEARN_SERVER:
140 protocols:
141 seldon:
142 defaultImageVersion: "1.19.0-dev"
143 image: seldonio/sklearnserver
144 v2:
145 defaultImageVersion: "1.3.5-sklearn"
146 image: seldonio/mlserver
147 TENSORFLOW_SERVER:
148 protocols:
149 seldon:
150 defaultImageVersion: "1.19.0-dev"
151 image: seldonio/tfserving-proxy
152 tensorflow:
153 defaultImageVersion: 2.1.0
154 image: tensorflow/serving
155 XGBOOST_SERVER:
156 protocols:
157 seldon:
158 defaultImageVersion: "1.19.0-dev"
159 image: seldonio/xgboostserver
160 v2:
161 defaultImageVersion: "1.3.5-xgboost"
162 image: seldonio/mlserver
163 TRITON_SERVER:
164 protocols:
165 v2:
166 defaultImageVersion: "21.08-py3"
167 image: nvcr.io/nvidia/tritonserver
168 HUGGINGFACE_SERVER:
169 protocols:
170 v2:
171 defaultImageVersion: "1.3.5-huggingface"
172 image: seldonio/mlserver
173 TEMPO_SERVER:
174 protocols:
175 v2:
176 defaultImageVersion: "1.3.5-slim"
177 image: seldonio/mlserver
178
179# ## Other
180# You can choose the crds to not be installed if you already installed them
181# This applies to just the yaml template. If you set managerCreateResources=true then
182# it will try to create the CRD but only if it does not exist
183crd:
184 create: true
185 # Whether to force the use of the v1beta1 or v1 CRD.
186 forceV1: false
187 forceV1beta1: false
188 annotations: {} # Annotations to add to the CRD if not created by the Manager
189
190# Warning: credentials will be depricated soon, please use defaultEnvSecretRefName above
191# For more info please check the documentation
192credentials:
193 gcs:
194 gcsCredentialFileName: gcloud-application-credentials.json
195 s3:
196 s3AccessKeyIDName: awsAccessKeyID
197 s3SecretAccessKeyName: awsSecretAccessKey
198
199kubeflow: false
200
201# Explainer image
202explainer:
203 image: seldonio/alibiexplainer:1.19.0-dev
204 image_v2: seldonio/mlserver:1.3.5-alibi-explain