Skip to main content

🌐 Deploying the Traffic Distributor via SaveinCloud Marketplace

📖 Introduction​

The Traffic Distributor is an intelligent load balancing solution available in the SaveinCloud marketplace.
It is designed to distribute HTTP/HTTPS traffic between two environments (backends) according to the configured percentage, ensuring scalability, high availability, and flexibility in application management.

🔎 What is the Traffic Distributor?​

The Traffic Distributor acts as a traffic balancer that allows you to:

  • ✅ Distribute requests proportionally between two backends.
  • ✅ Apply zero downtime updates (ZDT).
  • ✅ Perform A/B testing comparing application versions.
  • ✅ Ensure failover and high availability, maintaining active redundancy.

⚡ Benefits​

  • Zero Downtime Updates (ZDT): updates without interruption, directing all traffic to one backend while the other is under maintenance.
  • A/B Testing: routing part of the traffic to a new application version, enabling performance and user experience measurement.
  • High Availability: load distribution between two functional application instances, ensuring service continuity.

🛠️ Deploy via SaveinCloud Marketplace​

  1. Access the SaveinCloud Panel and go to the Marketplace tab.
    imagem botão do marketplace

  2. Search for the Traffic Distributor item and click Install.
    Traffic distributor no marketplace

  3. Configure the deployment parameters described below.
    Configurações pre-provisionamento Traffic Distributor

⚙️ Available Settings​

When installing the Traffic Distributor, the following fields must be configured:

  • Entrypoint
    Defines how external traffic will be received:

    • Shared LB → uses the platform's shared load balancer.
    • Public IP → assigns a dedicated public IP to the Traffic Distributor.

  • Balancers
    Number of load balancer instances to be provisioned for the service.

    💡 It is recommended to start with 1 and increase according to demand.

  • Routing Method
    Defines how traffic will be distributed between the backends:

    • Round Robin → distributes requests sequentially and evenly.
    • Sticky Sessions → keeps the same client connected to the same backend.
    • Failover → sends all traffic to a primary backend, switching to the secondary only in case of failure.

  • Traffic Ratio
    Allows defining the percentage of traffic sent to each backend.

    Example: 50/50 (equal distribution) or 70/30 (majority to one environment).

  • HTTPS
    Enables or disables secure traffic routing (HTTPS).

    • Requires backends to have valid certificates configured.

  • Backends
    Domains or IP addresses of the environments that will receive the traffic.

    • At least two destinations must be specified for load balancing.

  • Environment
    Technical identifier of the created environment, provided by the platform.

  • Display Name
    Friendly name to identify the service in the SaveinCloud panel.

  • Region
    Geographic location where the Traffic Distributor will be provisioned (Brasília, São Paulo, or Fortaleza).

📊 Use Cases​

  1. 🔄 Zero Downtime Updates (ZDT):
    During updates, direct 100% of the traffic to one backend while the other is updated.

  2. 🧪 A/B Testing:
    Configure 80% of the traffic to the current version and 20% to the new version, monitoring performance and UX.

  3. 🛡️ High Availability and Failover:
    Keep two active environments to ensure redundancy and fault tolerance.

✅ Best Practices​

  • Monitor traffic and error metrics to adjust distribution.
  • Apply A/B testing gradually before moving 100% of the traffic.
  • Use ZDT mode in critical updates to avoid downtime.
  • Combine with valid SSL certificates for HTTPS traffic security.

🐞 Troubleshooting​

  • Traffic is not being distributed correctly
    Check the percentage configuration.

    Confirm that the backends are active and accessible.

  • SSL/HTTPS error

    Ensure certificates are correctly applied on the backends.

  • One backend is not receiving traffic

    Verify that the defined ratio is not 100% for only one backend.
    Manually test connectivity with the backend.