Managed Load Balancer

UpCloud’s Managed Load Balancer allows users to create a fixed point of entry for services and applications that is able to distribute incoming connections according to a number of selectable balancing algorithms. Using the Managed Load Balancer enables users to build resilience and increase the scalability of web applications with an easy setup and flexible configuration.

Managed Load Balancer can be configured using the intuitive UpCloud Control Panel or by utilising the fully-featured UpCloud API. As a managed service, Load Balancer takes care of maintenance and software configuration creating a true turn-key service.

Balancing Algorithms

The load balancing algorithm determines the method according to which traffic to the service is distributed between backends servers. Users can choose between the following options:

• Round robin uses each backend server evenly in rotation. one of the simplest and most used
• Least connections direct traffic dynamically to the backend server with the lowest number of active connections. This method can be beneficial with longer-lived connections avoiding overloading any individual server.
• Source IP hashes the IP address of the incoming connection and divides the traffic evenly while ensuring the same backend is always chosen for the same source IP. This can be useful when it is important that a client should continue a session that might still be active after a disconnection.
• Hostname matching routes incoming traffic based on the domain name specified in the Host header allowing you to distribute traffic to e.g. example.com and blog.example.com to different backends.

Plans

Load Balancer offers two plans to support developers and large businesses alike. The service can be resized as needed to both increase and decrease resources.

Development plan

The Development plan allows users to test the service and begin developing applications without large upfront investments.

Supporting up to 1000 concurrent sessions, the Development plan consists of a single node for handling small to medium workloads, suitable for low traffic applications and development environments.

Production plan

Applications and services that need higher capacity or want to enable high availability can scale up to the Production plan.

The Production plan includes two parallel load balancing nodes both capable of handling up to 50000 sessions. Forming a high availability cluster, the Production plan Load Balancer distributes workloads across all nodes in order to prevent any single node from getting overloaded.

SSL Certificates

Managed Load Balancer service includes a built-in SSL certificate management system that enables users to dynamically obtain certificates or manually upload existing ones. Including an SSL certificate bundle to a Load Balancer frontend allows users to enable a secure connection via HTTPS.

These features are available at the UpCloud Control Panel as well as via the UpCloud API.

Manual

The Manual mode lets users import existing SSL certificate bundles.

Users are able to upload their SSL certificate, any included intermediaries, and the corresponding private key to create certificate bundles. These are usually obtained together via the service used to generate and authenticate the certificates.

Dynamic

The Dynamic mode sets a request to generate a new SSL certificate when attached to a Load Balancer frontend.

Creating a dynamic SSL certificate bundle allows users to obtain certificates directly to their Load Balancer by simply creating the necessary DNS records to demonstrate their authority over the claimed domain names.

Network Transfer Pool

Network Transfer Pool combines all transfer quotas included in the user’s UpCloud products and services into a single account-wide transfer pool, free of charge.

Load Balancer instances fully benefit from all network transfer capacity available in the user’s Network Transfer Pool. Users can track network quota and total usage at their UpCloud Control Panel or via the UpCloud API.

Maintenance

Any changes or updates are performed as rolling updates, meaning zero downtime for Production plans which use multiple nodes.