メインコンテンツまでスキップ
バージョン: 3.7 (unsupported)

Getting Started with Helm Charts (Scalar Manager)

Scalar Manager is a web-based dashboard that allows users to:

  • check the health of the Scalar products
  • pause and unpause the Scalar products to backup or restore underlying databases
  • check the metrics and logs of the Scalar products through Grafana dashboards

The users can pause or unpause Scalar products through Scalar Manager to backup or restore the underlying databases.
Scalar Manager also embeds Grafana explorers by which the users can review the metrics or logs of the Scalar products.

Assumption

This guide assumes that the users are aware of how to deploy Scalar products with the monitoring and logging tools to a Kubernetes cluster. If not, please start with Getting Started with Scalar Helm Charts before this guide.

Requirement

  • You need privileges to pull the Scalar Manager container (scalar-manager) from GitHub Packages.
  • You must create a Github Personal Access Token (PAT) with read:packages scope according to the GitHub document to pull the above container.

What we create

We will deploy the following components on a Kubernetes cluster as follows.

+--------------------------------------------------------------------------------------------------+
| +----------------------+ |
| | scalar-manager | |
| | | |
| | +------------------+ | --------------------------(Manage)--------------------------+ |
| | | Scalar Manager | | | |
| | +------------------+ | | |
| +--+-------------------+ | |
| | | |
| +------------------------------------+ | |
| | loki-stack | V |
| | | +-----------------+ |
| | +--------------+ +--------------+ | <----------------(Log)--------------- | Scalar Products | |
| | | Loki | | Promtail | | | | |
| | +--------------+ +--------------+ | | +-----------+ | |
| +------------------------------------+ | | ScalarDB | | |
| | | +-----------+ | |
| +------------------------------------------------------+ | | |
| | kube-prometheus-stack | | +-----------+ | |
| | | | | ScalarDL | | |
| | +--------------+ +--------------+ +--------------+ | -----(Monitor)----> | +-----------+ | |
| | | Prometheus | | Alertmanager | | Grafana | | +-----------------+ |
| | +-------+------+ +------+-------+ +------+-------+ | |
| | | | | | |
| | +----------------+-----------------+ | |
| | | | |
| +--------------------------+---------------------------+ |
| | | |
| | | Kubernetes |
+----+-----------------------+---------------------------------------------------------------------+
| |
expose to localhost (127.0.0.1) or use load balancer etc to access
| |
(Access Dashboard through HTTP)
| |
+----+----+ +----+----+
| Browser | <-(Embed)-- + Browser |
+---------+ +---------+

Step 1. Upgrade the kube-prometheus-stack to allow Grafana to be embedded

  1. Add or revise this value to the custom values file (e.g. scalar-prometheus-custom-values.yaml) of the kube-prometheus-stack

    grafana:
    grafana.ini:
    security:
    allow_embedding: true
    cookie_samesite: disabled
  2. Upgrade the Helm installation

    helm upgrade scalar-monitoring prometheus-community/kube-prometheus-stack -n monitoring -f scalar-prometheus-custom-values.yaml

Step 2. Prepare a custom values file for Scalar Manager

  1. Get the sample file scalar-manager-custom-values.yaml for scalar-manager.

  2. Add the targets that you would like to manage. For example, if we want to manage a ledger cluster, then we can add the values as follows.

    scalarManager:
    targets:
    - name: my-ledgers-cluster
    adminSrv: _scalardl-admin._tcp.scalardl-headless.default.svc.cluster.local
    databaseType: cassandra

    Note: the adminSrv is the DNS Service URL that returns SRV record of pods. Kubernetes creates this URL for the named port of the headless service of the Scalar product. The format is _{port name}._{protocol}.{service name}.{namespace}.svc.{cluster domain name}

  3. Set the Grafana URL. For example, if your Grafana of the kube-prometheus-stack is exposed in localhost:3000, then we can set it as follows.

    scalarManager:
    grafanaUrl: "http://localhost:3000"
  4. Set the refresh interval that Scalar Manager checks the status of the products. The default value is 30 seconds, but we can change it like:

    scalarManager:
    refreshInterval: 60 # one minute
  5. Set the service type to access Scalar Manager. The default value is ClusterIP, but if we access using the minikube tunnel command or some load balancer, we can set it as LoadBalancer.

    service:
    type: LoadBalancer

Step 3. Deploy scalar-manager

  1. Create a secret resource reg-docker-secrets to pull the Scalar Manager container image from GitHub Packages.

    kubectl create secret docker-registry reg-docker-secrets --docker-server=ghcr.io --docker-username=<github-username> --docker-password=<github-personal-access-token>
  2. Deploy the scalar-manager Helm Chart.

    helm install scalar-manager scalar-labs/scalar-manager -f scalar-manager-custom-values.yaml

Step 4. Access Scalar Manager

If you use minikube

  1. To expose Scalar Manager's service resource as your localhost (127.0.0.1), open another terminal, and run the minikube tunnel command.

    minikube tunnel
  2. Open the browser with URL http://localhost:8000

If you use other Kubernetes than minikube

If you use a Kubernetes cluster other than minikube, you need to access the LoadBalancer service according to the manner of each Kubernetes cluster. For example, using a Load Balancer provided by cloud service or the kubectl port-forward command.

Step 5. Delete Scalar Manager

  1. Uninstall scalar-manager
    helm uninstall scalar-manager