Cosmo Tech tenant

install Cosmo Tech API and all its dependencies in a dedicated namespace

Requirements

• working Kubernetes cluster deployed from Cosmo Tech terraform-provider (like terraform-azure for example)
• terraform

  If using Windows, Terraform must be accessible from PATH

• situational
  • Azure: azure-cli
  • AWS: aws-cli

How to

• clone & open the repository

  git clone https://github.com/Cosmo-Tech/terraform-tenant.git --branch <tag>
  cd terraform-tenant
  

• deploy
  • fill terraform.tfvars variables according to your needs
  • run pre-configured script

    ℹ️ comment/uncomment the terraform apply line at the end to get a plan without deploy anything

    • Linux

      ./_run-terraform.sh
      

    • Windows

      ./_run-terraform.ps1
      

  • Azure
    • will ask for the access key of the Azure Storage of "cosmotechstates"
      • go to Azure > Azure Storage > "cosmotechstates" > Access keys
      • copy/paste "Key" from "key1" or "key2" in the terraform input
  • AWS
    • to fill
  • GCP
    • to fill

Known errors

• Error: Provider configuration not present

  The tenant has been deleted or exists in an other cluster, but the state file still exists. Chose another tenant name or delete the state file if tenant doesn't exist anymore.

• Error: [POST /datasources][409] addDataSourceConflict {"message":"data source with the same name already exists"} [..] with module.config_grafana_dashboard.grafana_data_source.redis-datasource

  The state file could not be found (it has probably been deleted, but the deployed resources remains) Go to https://<cluster_url>/monitoring (credentials are stored on Kubernetes secret monitoring/kube-prometheus-stack-grafana) -> Connections -> Data sources -> Delete both tenant--postgresql and tenant--redis

• Error: failed to create folder: [POST /folders][409] createFolderConflict {"message":"a folder with the same name already exists in the current location"} [..] with module.config_grafana_dashboard.grafana_folder.folder

  The state file could not be found (it has probably been deleted, but the deployed resources remains) Go to https://<cluster_url>/monitoring (credentials are stored on Kubernetes secret monitoring/kube-prometheus-stack-grafana) -> Dashboards -> Delete tenant-

• Error: error sending POST request to /keycloak//admin/realms: 409 Conflict. Response body: {"errorMessage":"Conflict detected. See logs for details"} [..] with with module.config_keycloak_realm.keycloak_realm.realm

  The state file could not be found (it has probably been deleted, but the deployed resources remains) Go to https://<cluster_url>/keycloak (credentials are stored on Kubernetes secret keycloak/keycloak-config) -> Select the realm "tenant-" -> Realm settings > Action > Delete

Developpers

• modules
  • terraform-tenant
    • chart_argo = install Argo Workflows
    • chart_cosmotech_api = install Cosmo Tech API
    • chart_postgresql = install PostgreSQL (and configure it for Cosmo Tech API, SeaweedFS & Argo Workflows)
    • chart_redis = install Redis
    • chart_seaweedfs = install SeaweedFS
    • config_grafana_dashboard = create tenant configuration on existing Grafana instance
    • config_keycloak_realm = create tenant configuration on existing Keycloak instance
    • kube_namespace = create tenant namespace
    • storage = [temporary] dynamically create persistence storage for charts requiring it
• Terraform state
  • The state is stored beside the cluster Terraform state, in the current cloud s3/blob storage service (generally called cosmotech-states or cosmotechstates, depending on what the cloud provider allows in naming convention)
• File backend.tf
  • dynamically created at each run of _run-terraform
  • permit to have multi-cloud compatibility with Terraform
  • it instanciate the needed Terraform providers based on the variable cloud_provider from terraform.tfvars
  • this file is a workaround to avoid having unwanted variables related to cloud providers not targetted in current deployment

Made with ❤️ by Cosmo Tech DevOps team