Dsquares Loyalty Platform
  • Overview
  • App Components
    • Application Components Diagrams
    • API Gateway
    • Users Service
    • Merchants Service
    • Pointing Service
    • Vouchers Service
    • Payment Service
    • Channels Service
    • Gamification
    • C-Cubed
  • Hosting and Network
    • High Level Infrastructure Design
    • Hardware Requirements
      • Cloud
      • Bare-Metal Installation
      • Kubernetes
    • High Availability & Disaster Recovery
    • Observability
    • On-Premise Components
      • Introduction
      • Most workload on Dsquares cloud
      • Most workload on customer's premise
      • Hybrid setup
    • KSA Based Partners Solution
  • Data Flow
    • Description
    • Sequence Diagrams
      • DSquares System
      • Data Warehouse
      • Gamification
      • C-Cubed
      • API Gateway
      • Pointing Service
      • Vouchers service
      • Payment Service
      • Channels Service
  • Integrations
    • Introduction
    • Data Pipelines
      • Messaging Queue
        • Dsquares Kafka Library
    • Rest APIs
      • Rewards
        • Get Deals
        • Subscribe to offer
        • Verify Coupon Number
        • Burn Voucher
        • Burn PreGenerated Coupon
      • Points
        • Reward
        • Refund
        • Redeem User
        • Cancel Points
        • Get Wallet Data
      • 🎮Gamification
        • GetChallengesInfoForUser
        • GetProgressChallengesInfoForUser
        • GetBadgeChallenges
        • GetChallengeHistoryForUser
        • GetDetailsById
        • GetBadgeDetailsById
        • GetChallengeCategories
        • AddAction
        • UpdateIsClaimed
      • InsightsGathering
        • Participations Insights
      • Errors
  • Security
    • ISO 27001
    • Authentication
Powered by GitBook
On this page

Cloud

PreviousHardware RequirementsNextBare-Metal Installation

Last updated 8 months ago

  • Kubernetes Cluster, should be highly available with two worker nodes in two different AZs at least (or three in case of installing Kafka either using Zookeeper or KRaft inside the cluster since it requires at least two replicas running).

  • Node sizing depends on which other services will be managed or included in the cluster and on the load on the services, but a good start would be two worker nodes 4 CPUs, 16 GB, 100 GB HD RAM each.

  • We will need about 25 GB monthly for logs and metrics.

  • If available, managed services for:

    1. Redis (Instance, cluster in the future) (for caching), we recommend 1 vCPU, 1 GB RAM

    2. MSSQL (Highly Available one master and at least one synchronous replica, with license), we recommend 6 vCPUs, 32 GB RAM, 500 GB SSD

    3. MySQL (Highly Available, one master and at least one synchronous replica), we recommend 2 vCPUs, 8 GB RAM, 500 GB SSD

    4. MongoDB (Highly Available, 3 replicasets with replication factor of at least 2), we recommend 4vCPUs, 16 GB RAM, 80 GB SSD

    5. PostgreSQL (Highly Available), for operational workload (keycloak, vault, etc...) if needed

    6. Kafka Cluster (Highly Available, 3 replicas with replication factor of at least 2) 2 vCPUs, 8 GB RAM, 100 GB PV (high performance) each

    7. RabbitMQ

  • Network Services for:

    1. VPC with private subnet for all servers (Kubernetes worker nodes and windows servers)

    2. Application Load Balancer (for Windows Server Machines)

    3. NAT Gateway

Page cover image
GCP Setup as an Example