Kubernetes - Configuring values.yaml

The values.yaml file is the core configuration file for your ScopeAR deployment.  It contains various settings you can customize to fit your environment, such as cluster information, replica counts, logging levels, and resource limits.

Below is a step-by-step guide to each key in the file, explaining its functionality.

Cluster and Environment Settings

These settings control the overall deployment environment for your Kubernetes cluster. If you would like a sample values.yaml to start with, you can find one at scopearctl/docs/sample/scopear_k8s_eks_sample.yaml.

• metadata.name: Name of the Kubernetes cluster. For example, eks-dev-0.

• metadata.namespace: The namespace for the ScopeAR deployment defaults to 'scopear'

• spec.env: Specify 'intranet' to include client apps in CMS. If apps aren’t needed, use ‘production’

• spec.replicas: Global replica count for all services (default: 1). This can be overridden if specified for the individual service.

• spec.logLevel: Global log level for services (default: info)

• subdomainSuffix: Suffix appended to subdomains (e.g., cms-suffix.example.com)

• hostnames: Combined with subdomainSuffix for the entire domain (e.g., cms-suffix.<primary domain>.com)

Image Registry Configuration

These settings control where Docker images are pulled from.

• imageRegistry: Specify the internal image registry, if applicable. Defaults to http://docker.io

• stripPublicRegistry: If true, strips 'Docker: Accelerated Container Application Development ' from image names. Defaults to false.

Cluster Provider Configuration

Define the cluster provider and other platform-specific settings.

• cluster.provider: Cluster provider. It can be one of aws, google, azure, or other.

• cluster.type: Cluster type. It can be one of eks, kops,gke, aks, or openshift.

CMS Service Configuration

Settings specific to the CMS service.

• cms.eksRoleArn: AWS Role ARN with S3 GetObject, PutObject, and ListBucket permissions for the CMS S3 bucket, as defined in storage.default_bucket. 

• cms.config.admin.email_address: The email address of the super user admin. This email address will be the first user seeded in the CMS. Please contact Scope AR for default deployment password.

• cms.config.mysql.host: MySQL hostname. If using in-cluster MySQL, this can be ignored

• cms.config.mysql.use_iam_auth: Boolean for using IAM auth for connection to AWS RDS. Defaults to false

• cms.config.aws_rds_region: AWS region where RDS instance is located. It can be ignored if using in-cluster MySQL

• cms.config.redis.hostname: Redis hostname. If using in-cluster Redis, this can be ignored

• cms.config.smtp.smtp_authentication: Specify the authentication type here (plain/login/cram_md5) - default: plain

• cms.config.smtp.smtp_openssl_verify_mode: If using TLS, set how OpenSSL checks the certificate (none/peer/client_once/fail_if_no_peer_cert). Defaults to 'none'

• cms.config.smtp.smtp_enable_starttls_auto: Defaults to ‘true’ for SMTP using TLS

• cms.config.storage.file_storage: Can be one of ‘s3/minio/local’. It is recommended that ‘minio’ be used for local storage. This allows Worklink Create and CMS to use the same storage backend. If set to local

• cms.config.storage.storage_ssl_verify: If using AWS services or creating trusted SSL certificates for Minio, set to true. If using self-signed certificates for Minio, set to false.

• cms.config.storage.default_bucket: Name of S3/Minio bucket to use for content storage.

• cms.config.storage.default_region: If using S3, the default region where S3 bucket is located.

• cms.config.zeus.enabled: If using remote assist feature, enable zeus.

C2 Service Configuration

This section configures the C2 service, including API, client, project, packager, and model converter settings.

• c2.eksRoleArn: AWS Role ARN with S3 GetObject, PutObject, and ListBucket permissions for the C2 S3 bucket, as defined in config.storage.bucket_name. 

• c2.api, client, model_converter, project_packager: Set appropriate resource request/limits

• c2.config.cad_pipeline.processing_mode: Defaults to ‘none’. Processing mode for model conversion. This can be one of ‘pixyz_local, pixyz_azure, none.’ If Pixyz licenses have been purchased, please contact support@scopear.com for further instructions.

• c2.config.cad_pipeline.flex_host: FlexLM license server hostname. Only applicable if utilizing Pixyz for model conversion.

• c2.config.cad_pipeline.flex_port: FlexLM license server port.

• c2.config.mongodb.database: Database and connection string to connect to MongoDB. If using DocumentDB, please use ‘ssl=true’ in your connection string (i.e. scopear?ssl=true&retryWrites=false&tlsInsecure=true). If using in-cluster mongodb, set ‘ssl=false’ in connection string (i.e. scopear?ssl=false&retryWrites=false&tlsInsecure=true)

• c2.config.storage.storage_provider: Can be one of ‘s3’ or ‘minio’

• c2.config.storage.bucket_name: Bucket name for minio/s3

Supporting Services

This section configures the supporting services for CMS/C2, which includes Zeus, Diagnostic Tools, and IoT.

Zeus (Presence Server)

• zeus.annotations: Zeus allows you to add annotations if you are deploying to a cloud provider and would like to use annotations to provision the NLB. The annotations provided above are only an example. Please use annotations that fit your cloud provider. If you are provisioning a load balancer outside annotations, please ignore this section

• zeus.config.redis.database: The Zeus Redis database number. Defaults to ‘10’

• zeus.config.redis.use_auth: Whether to specify an AWS token for authentication. Default to ‘false’

Diagnostic Tools

This section configures the diagnostic tools server. The only options are to enable it and specify replicas. We recommend leaving this to 1 replica since it is an internal testing tool only.

IoT Server

This section configures the IoT server. This is an optional package and requires more in-depth scoping before enabling. Please contact support@scopear.com if you are interested or have any questions.

• iot.annotations: IoT allows you to add annotations if you are deploying to a cloud provider and would like to use annotations to provision the NLB. The annotations provided above are only an example. Please use annotations that fit your cloud provider. If you are provisioning a load balancer outside annotations, please ignore this section.

In-Cluster Datastores

If your organization cannot leverage cloud services for Redis/s3/MySQL/MongoDB, you will need to configure the data stores within the cluster. Kubernetes supports dynamic provisioning of storage volumes, which means that when you create a PersistentVolumeClaim (PVC), Kubernetes can automatically provision the underlying storage resource (e.g., a disk in a public cloud, an NFS share, a Ceph RBD volume, etc.). However, if you cannot leverage dynamic provisioning and need to create the volumes manually beforehand, we support specifying volume names in your volume claim templates.

Dynamic Provisioning

Below is an example of deploying in-cluster data stores, leveraging cloud-based storage provisioners. This method is typical when the admin wants to deploy to EKS, AKS, GKE, dynamically, utilizing the default storage controller/storage class name configured in the cluster. If leveraging external datastores like S3, RDS, DocumentDB, please set enabled: false

Additionally, if S3 is not a storage option, and you are migrating from a previous installation where local storage was used for CMS, you will need to specify the following:

• cms.config.storage.storage_size: Size of the disk to be mounted in CMS pod.

• cms.config.storage.volume_name: If pre-built disks are desired, specify the volume name.

• cms.config.storage.volume_mode: Volume mode of the disk. If not specified, this will use the default volume mode of storage provisioned. (e.g. Block, Filesystem, etc)

• cms.config.storage_class_name: Storage class name to utilize. If not specified, will use the default storage class name of the provider. (e.g. gp2, nfs-csi, standard, etc)

Non-Dynamic Provisioning

• storage_size: Even when specifying volume_name of a pre-configured disk, you must specify the desired storage size, making sure it is >= the actual size of the pre-built volume.

• storage_class_name: Storage class name to utilize. If not specified, will use the default storage class name of the provider. (e.g. gp2, nfs-csi, standard, etc)

• volume_mode: Volume mode of the disk. If not specified, this will use the default volume mode of storage provisioned. (e.g. Block, Filesystem, etc)

• volume_name:If pre-built disks are desired, specify the volume name.

Full Example - AWS EKS

Below is an example values.yaml configuration for an AWS EKS deployment, utilizing AWS Services for the datastores with Pixyz and IoT enabled:

Full Example - In-Cluster Data Stores

Below is an example of deploying Scope AR into an on-prem Openshift cluster with in-cluster data stores (minio/mysql/mongodb/redis):