Getting Started

Recommendations for selecting upstreams

When selecting an upstream server, it is recommended to use the server's private IP address, as traffic routed through the GPN is free of charge. If a public IP address is used, the traffic will be routed through the internet, incurring additional charges.

To create a load balancer, you need at least one provisioned server.

In the left-hand menu, navigate to Load Balancers and click Create new balancer
Select the type of balancer. See Types of load balancing
Choose one of the available locations

Configure the load balancer settings
The settings are divided into sections in the interface. Fill in the fields in each section

Upstream groups – Here you can configure the Upstream groups

Setting	Description
Group name	The name of the pool of upstream servers
Protocol	The protocol used by the load balancer to communicate with upstream servers (for HTTPS L7 load balancer): HTTP(S) - supports standard HTTP/HTTPS for web requests and requires an SSL certificate gRPC(S) - supports gRPC over HTTP/2 for high-performance service-to-service communication and requires an SSL certificate
Upstream server	Create the upstream servers pool by selecting servers from the dropdown list and clicking Add upstream server For each upstream server, specify: The Port to which the load balancer will forward requests for that server The Weight (priority) for distributing traffic (servers with higher weights will receive more requests)
Algorithm	Choose one of the available traffic distribution algorithms. See Algorithms
Sticky cookie	Tick the checkbox if you want to enable session persistence, binding all user requests to the same upstream server (for HTTPS L7 load balancer)

Health check settings – In this section, configure the health check parameters for the upstream servers

For HTTPS L7 load balancer
Setting	Description
Domain	The domain used in health check requests
Health check path	The path on the upstream server used to check availability
Request method	The HTTP method for health checks: GET - requests the full response from the server HEAD - requests only response headers, reducing server load
Checks to fail	The number of consecutive failed checks required to mark a server as unhealthy
Checks to pass	The number of consecutive successful checks required to mark a server as healthy
Inspect new upstream servers	If enabled, the load balancer performs health checks on new upstream servers before adding them to the active pool
For both load balancer types
Interval	The time in seconds between health checks for the upstream servers
Jitter	A random value to be added to the check interval, reducing network load and improving stability

Forwarding rules – Here you can configure forwarding rules

For both load balancer types
Setting	Description
Rule name	The name of the rule
Ports	The ports through which the load balancer will accept incoming requests
For HTTPS L7 load balancer
Domains	The domains handled by specified forwarding rule
Protocol	Defines the protocol for incoming request handling: HTTP - for unencrypted connections HTTPS - for encrypted connections using SSL/TLS HTTP/2 - if enabled, the load balancer supports requests using the HTTP/2 protocol
Redirect HTTP to HTTPS	If enabled, all HTTP requests are automatically redirected to HTTPS
Minimum TLS version	Sets the minimum allowed TLS version for HTTPS connections: TLSv1.2 - widely supported version with high compatibility TLSv1.3 - provides enhanced security and performance but may not be supported by older clients
SSL certificate	Allows selecting a pre-created SSL certificate to encrypt HTTPS connections on the load balancer

Rewrites – Enables path rewriting for incoming requests, redirecting them to a specified path and corresponding upstream server group (for HTTPS L7 load balancer)

Setting	Description
Location	The original path or URI to be rewritten
Upstream path	The new path or URI to which the request will be redirected
Upstream group	The upstream server group to which the request will be directed after path rewriting

Set real IP from header – Here you can configure the forwarding of the client's real IP address to the upstream server (for HTTPS L7 load balancer)

Setting	Description
Disabled	The real IP address is not forwarded; upstream servers see the load balancer's IP address
X-Real-IP	Uses the `X-Real-IP` header to forward the client's real IP address
X-Forwarded-For	Uses the `X-Forwarded-For` header, where the client's real IP address is listed first

Custom error pages – Here you can configure custom error pages for specific HTTP codes (for HTTPS L7 load balancer)

Setting	Description
Error code	Displays a custom page when a specified HTTP error code occurs
Redefine HTTP code	Replaces the displayed HTTP code with another (e.g., change 500 to 404)
Upload file	Uploads a custom error page file (maximum size: 4096 bytes)
File name	The name of the uploaded file
File size	The size of the uploaded file
Content type	Selects the content type of the uploaded file (e.g., `text/html`)

Ports – Here you can configure ports (for TCP L4 Load Balancer)

Setting	Description
Proxy Protocol	Enables forwarding of the client's real IP address to upstream servers. When enabled, upstream servers see the client's original IP instead of the load balancer's IP
Upstream Group	Selects the upstream server group to which requests will be forwarded

Additional options – Here you can configure additional options

Setting	Description
Geo IP	If enabled, the load balancer determines the client's geographical location based on their IP address. This information can be used for request routing or access restrictions (for HTTPS L7 load balancer)
Logs Storing	Tick the checkbox if you want to save logs of requests processed by the load balancer. Logs will be stored in the cloud storage in NGINX format

In the Cluster section, specify the name of the dedicated cluster (if applicable)
In the Name section, provide a name for the load balancer to distinguish it from others in the list
Click Create