Helm Values

Schema for the values.yaml file for the tensor-fusion Helm chart

Property	Type	Constraints	Description
imagePullSecrets ↓	array		List of secrets for pulling images from private repositories. See: https://kubernetes.io/docs/tasks/configure-pod-container/pull-image-private-registry/
nameOverride	string		Override for the chart name
fullnameOverride	string		Override for the full name of resources created by this chart
namespaceOverride	string		Override for the namespace where resources will be deployed
serviceAccount ↓	object		Configuration for the service account. See: https://kubernetes.io/docs/concepts/security/service-accounts/
initialGpuNodeLabelSelector	string		Label selector used to identify GPU nodes in the cluster Default: nvidia.com/gpu.present=true
controller ↓	object		Configuration for the tensor-fusion controller component
greptime ↓	object		Configuration for GreptimeDB integration
agent ↓	object		Configuration for the tensor-fusion agent component
cloudVendorCredentials ↓	object		Cloud vendor credentials for pools running in Provisioned mode without IRSA or zero-credential auth
alert ↓	object		Configuration for alerting
dynamicConfig ↓	object		Dynamic configuration for alerting

imagePullSecrets (items)

List of secrets for pulling images from private repositories. See: https://kubernetes.io/docs/tasks/configure-pod-container/pull-image-private-registry/

Properties

Property	Type	Constraints	Description
name	string		Name of the secret containing the registry credentials

serviceAccount

Configuration for the service account. See: https://kubernetes.io/docs/concepts/security/service-accounts/

Properties

Property	Type	Constraints	Description
create	boolean		Specifies whether a service account should be created Default: true
automount	boolean		Automatically mount a ServiceAccount's API credentials Default: true
annotations	object		Annotations to add to the service account

controller

Configuration for the tensor-fusion controller component

Properties

Property	Type	Constraints	Description
replicaCount	integer	min: 1	Number of replicas for the controller deployment Default: 1
image ↓	object		Container image configuration for the controller
vectorAgentImage	string		Tag for the vector agent, for collecting metrics to TSDB Default: docker.io/timberio/vector:latest-alpine
podAnnotations	object		Annotations to add to the controller pods
tolerations	array		Tolerations for the controller pods
affinity	object		Affinity rules for the controller pods
livenessProbe	object		Liveness probe configuration for the controller pods
readinessProbe	object		Readiness probe configuration for the controller pods
resources ↓	object		Resource requirements for the controller
admissionWebhooks ↓	object		Configuration for the admission webhooks

image

Container image configuration for the controller

Properties

Property	Type	Constraints	Description
repository	string		Repository for the controller image Default: tensorfusion/tensor-fusion-operator
tag	string		Tag for the controller image Default: latest

resources

Resource requirements for the controller

Properties

Property	Type	Constraints	Description
requests ↓	object		Resource requests for the controller
limits ↓	object		Resource limits for the controller

requests

Resource requests for the controller

Properties

Property	Type	Constraints	Description
cpu	string		CPU request for the controller
memory	string		Memory request for the controller

limits

Resource limits for the controller

Properties

Property	Type	Constraints	Description
cpu	string		CPU limit for the controller
memory	string		Memory limit for the controller

admissionWebhooks

Configuration for the admission webhooks

Properties

Property	Type	Constraints	Description
failurePolicy	string	Fail Ignore	Failure policy for the admission webhooks Default: Fail
secretName	string		Name of the secret containing the webhook certificates Default: tensor-fusion-webhook-secret
patch ↓	object		Configuration for patching the webhook certificates

patch

Configuration for patching the webhook certificates

Properties

Property	Type	Constraints	Description
image	string		Image for the webhook certificate generator Default: registry.k8s.io/ingress-nginx/kube-webhook-certgen:v1.5.0

greptime

Configuration for GreptimeDB integration

Properties

Property	Type	Constraints	Description
isCloud	boolean		Whether to use GreptimeDB Cloud Default: false
host	string		Hostname of the GreptimeDB server Default: greptimedb-standalone.greptimedb.svc.cluster.local
port	integer		Port of the GreptimeDB server Default: 4001
image ↓	object		Container image configuration for GreptimeDB
resources ↓	object		Resource requirements for GreptimeDB
installStandalone	boolean		Whether to install a standalone GreptimeDB instance Default: true
user	string		Username for GreptimeDB authentication
db	string		Database name in GreptimeDB
password	string		Password for GreptimeDB authentication

image

Container image configuration for GreptimeDB

Properties

Property	Type	Constraints	Description
repository	string		Repository for the GreptimeDB image, for China mainland users, should change 'greptime.image.repository' value to 'greptime-registry.cn-hangzhou.cr.aliyuncs.com/greptime/greptimedb' Default: docker.io/greptime/greptimedb
tag	string		Tag for the GreptimeDB image Default: latest

resources

Resource requirements for GreptimeDB

Properties

Property	Type	Constraints	Description
requests ↓	object		Resource requests for GreptimeDB
limits ↓	object		Resource limits for GreptimeDB

requests

Resource requests for GreptimeDB

Properties

Property	Type	Constraints	Description
cpu	string		CPU request for GreptimeDB
memory	string		Memory request for GreptimeDB

limits

Resource limits for GreptimeDB

Properties

Property	Type	Constraints	Description
cpu	string		CPU limit for GreptimeDB
memory	string		Memory limit for GreptimeDB

agent

Configuration for the tensor-fusion agent component

Properties

Property	Type	Constraints	Description
enrollToken	string		Token used for agent enrollment with the cloud service Default: token-from-cloud
agentId	string		Unique identifier for the agent in the format 'org:env' Default: org-from-cloud:env
cloudEndpoint	string		WebSocket endpoint for cloud communication Default: wss://app.tensor-fusion.ai
image ↓	object		Container image configuration for the agent
resources ↓	object		Resource requirements for the agent

image

Container image configuration for the agent

Properties

Property	Type	Constraints	Description
repository	string		Repository for the agent image Default: tensorfusion/tensor-fusion-agent
tag	string		Tag for the agent image Default: latest

resources

Resource requirements for the agent

Properties

Property	Type	Constraints	Description
requests ↓	object		Resource requests for the agent
limits ↓	object		Resource limits for the cluster agent

requests

Resource requests for the agent

Properties

Property	Type	Constraints	Description
cpu	string		CPU request for the agent Default: 50m
memory	string		Memory request for the cluster agent Default: 64Mi

limits

Resource limits for the cluster agent

Properties

Property	Type	Constraints	Description
cpu	string		CPU limit for the cluster agent Default: 1000m
memory	string		Memory limit for the cluster agent Default: 512Mi

cloudVendorCredentials

Cloud vendor credentials for pools running in Provisioned mode without IRSA or zero-credential auth

Properties

Property	Type	Constraints	Description
accessKey	string		Access key for cloud vendor authentication Default: dummy
secretKey	string		Secret key for cloud vendor authentication Default: dummy

alert

Configuration for alerting

Properties

Property	Type	Constraints	Description
enabled	boolean		Enable alerting Default: false
image ↓	object		Image for alerting
replicaCount	integer		Number of replicas for alerting Default: 1
resources ↓	object		Resource requirements for alerting
alertManagerConfig	object		Alertmanager configuration

image

Image for alerting

Properties

Property	Type	Constraints	Description
repository	string		Repository for the alerting image Default: dummy
tag	string		Tag for the alerting image Default: dummy

resources

Resource requirements for alerting

Properties

Property	Type	Constraints	Description
requests ↓	object		Resource requests for alerting
limits ↓	object		Resource limits for alerting

requests

Resource requests for alerting

Properties

Property	Type	Constraints	Description
cpu	string		CPU request for alerting
memory	string		Memory request for alerting

limits

Resource limits for alerting

Properties

Property	Type	Constraints	Description
cpu	string		CPU limit for alerting
memory	string		Memory limit for alerting

dynamicConfig

Dynamic configuration for alerting

Properties

Property	Type	Constraints	Description
metricsTTL	string		Metrics and Logs Retention Period, default to 30d
metricsFormat	string		Metrics format, default to 'influx', could be 'json' or 'otel'
metricsExtraPodLabels	array		Extra pod labels to be added to metrics
alertRules ↓	array		Alerting rules

alertRules (items)

Alerting rules

Properties

Property	Type	Constraints	Description
name	string		Name of the alerting rule
query	string		Query for the alerting rule
threshold	number		Threshold for the alerting rule
severity	string		Severity for the alerting rule
evaluationInterval	string		Evaluation interval for the alerting rule
consecutiveCount	integer		Consecutive count for the alerting rule
summary	string		Summary for the alerting rule, can use $labels to access labels, eg. {{ .pool }} {{ .worker }} {{ .Threshold }}
description	string		Description for the alerting rule, can use $labels to access labels, eg. {{ .pool }} {{ .worker }} {{ .Threshold }}
alertTargetInstance	string		Alert target instance for the alerting rule