Skip to main content

☸️ Kubernetes Cluster Deployment on SaveInCloud Marketplace

The Kubernetes Cluster from the SaveInCloud Marketplace allows provisioning a dedicated cluster, optional observability modules, and different topologies, serving both development and production environments.

🚀 Deployment Step-by-Step​

  1. Access the SaveInCloud Marketplace.
    imagem botão do marketplace
  2. Search for Kubernetes Cluster.
    K8s Cluster no Marketplace
  3. Click Install.
  4. Configure the options as described below.
    Configurações K8s pré provisionamento

⚙️ Configuration Options​

🔢 Version​

Select the desired Kubernetes version (e.g., v1.31.3).
It is recommended to always use the latest stable version.

📊 K8s Dashboard​

  • v2 → Official native Kubernetes dashboard, with a simple interface and basic management features.
  • Skooner → A lighter and more responsive alternative, focused on monitoring and workload management, RBAC, and advanced visualization.

🏗️ Topology​

  • Development → Simplified cluster (1 Control Plane, 1 Storage, 1 group of Workers).
  • Production → Distributed and redundant cluster (multiple Control Planes, multiple Storages, multiple Workers, and API Balancers).

🔎 Practical example:

  • Dev → ideal for testing, development, and staging.
  • Prod → recommended for critical workloads, ensuring high availability.

🌐 Ingress Controller​

Defines how external traffic will be routed to cluster services:

  • NGINX → Default option, robust and widely supported.
  • HAProxy → Focused on high L4/L7 performance.
  • Traefik → Designed for microservices, with native Let's Encrypt integration.

📦 Deployment​

  • Clean Cluster → Basic cluster, ready to receive workloads.
  • Custom → Allows adjusting specific configurations (when available).

💾 NFS Storage​

Enables persistent storage for workloads.
⚠️ Required to enable monitoring and tracing modules.

📈 Optional Modules​

Available only if NFS Storage is enabled:

  • Prometheus & Grafana → Advanced monitoring and metrics dashboards.
  • Jaeger Tracing Tools → Observability for distributed applications, request tracing between microservices.
  • Remote API Access → Allows remote access to the cluster API for integration via kubectl or CI/CD pipelines. (Does not require NFS Storage)

🌍 Environment & Region​

  • Environment → Technical name of the cluster (auto-generated, can be customized).
  • Display Name → Friendly name shown on the dashboard.
  • Region → Choose from available datacenters.

🗺️ Dev vs Prod Differences​

📌 Dev Environment

  • 1 Control Plane
  • 1 Storage
  • 1 group of Workers

📌 Prod Environment

  • 2 or more API Balancers (HAProxy)
  • 3 Control Planes
  • 3 Storages (auto-clustering)
  • 2 or more groups of Workers

➡️ This ensures greater resilience and high availability.

📥 Provisioning​

  1. Define the desired version, dashboard, topology, and ingress.
  2. If needed, enable NFS Storage and select optional modules.
  3. Configure the environment name and region.
  4. Click Install.

⏳ The process takes a few minutes and, at the end, the cluster will be ready for use.

✅ Best Practices​

  • Use Dev only for testing; always choose Prod for critical applications.
  • Enable NFS Storage + Prometheus/Grafana for monitoring from the start.
  • Plan the use of Ingress Controller according to the application architecture.
  • Use Remote API Access to integrate CI/CD pipelines.

🛠️ Troubleshooting​

  • Cluster does not start → Check if the selected Kubernetes version is supported.
  • Ingress does not respond → Confirm if the domain is correctly configured for the Load Balancer.
  • Modules do not enable → Enable NFS Storage before selecting Prometheus/Grafana or Jaeger.
  • SSL → Can be configured later via ingress with Let's Encrypt support in the Add-on.