Creating load balancers

Limitations

The forwarding rule and protocol cannot be changed after the load balancer pool is added.
If an IPv6 subnet where a load balancer will operate works in the SLAAC or DHCPv6 stateless mode, the load balancer will receive an IPv6 address automatically.

Prerequisites

A network where a load balancer will operate has IP management enabled.
All VMs that will be added in balancing pools have fixed IP addresses.

To create a load balancer with balancing pools

On the Load balancers screen, click Create load balancer.
In the Create load balancer window, do the following:
1. Specify a name and, optionally, description.
  
  A description should not contain any personally identifiable information or sensitive business data.
2. Enable or disable high availability:
  - With high availability enabled, two load balancer instances will be created. They will work in the Active/Standby mode according to the Virtual Router Redundancy Protocol (VRRP).
  - With high availability disabled, a single load balancer instance will be created.
In the Network settings section, select the network that the load balancer will operate in and, optionally, specify an IP address that will be allocated to the load balancer.
- If you selected a virtual network that is connected to a physical network via a router
  
  In this case, you can assign a floating IP address to the load balancer. To do it, select Use a floating IP address, and then choose either to use an available floating IP address or to create a new one.
- If you selected a shared physical network with both IPv4 and IPv6 subnets
  
  In this case, you need to choose the IP version that will be used for the load balancer.
In the Balancing pools section, create a balancing pool to forward traffic from the load balancer to virtual machines by clicking Add.

In the Create balancing pool window that opens, do the following:
1. In the Forwarding rule section, select a forwarding rule from the load balancer to the backend protocol:
  - With the HTTPS -> HTTPS rule
    Specify ports for incoming and destination connections.
    
    Ensure that all virtual machines have the same SSL certificate (or a certificate chain).
    
    Enable the PROXY protocol version 1 to add a human-readable header with connection information (the source IP address, destination IP address, and port numbers) as a part of the request header.
  - With the HTTPS -> HTTP rule
    Specify ports for incoming and destination connections.
    
    Upload an SSL certificate (or a certificate chain) in the PEM format and a private key in the PEM format.
    
    Choose HTTP headers to insert into the request.
    
    Enable the TLS encryption to re-encrypt traffic from the load balancer to its members.
    
    Enable the PROXY protocol version 1 to add a human-readable header with connection information (the source IP address, destination IP address, and port numbers) as a part of the request header.
  - With the HTTP -> HTTP rule
    Specify ports for incoming and destination connections.
    
    Choose HTTP headers to insert into the request.
    
    Enable the TLS encryption to re-encrypt traffic from the load balancer to its members.
    
    Enable the PROXY protocol version 1 to add a human-readable header with connection information (source IP address, destination IP address, and port numbers) as a part of the request header.
  - With the TCP -> TCP rule
    Specify ports for incoming and destination connections.
    
    Enable the TLS encryption to re-encrypt traffic from the load balancer to its members.
  - With the UDP -> UDP rule
    
    Specify ports for incoming and destination connections.
2. In the Balancing settings section, select the balancing algorithm:
  - Least connections. Requests will be forwarded to the VM with the least number of active connections.
  - Round robin. All VMs will receive requests in the round-robin manner.
  - Source IP. Requests from a unique source IP address will be directed to the same VM.
  Enable/disable the Sticky session option to enable/disable session persistence. The load balancer will generate a cookie that will be inserted into each response. The cookie will be used to send future requests to the same VM.
  
  This option is not available in the SSL passthrough mode.
3. In the Members section, add members, that is, virtual machines, to the balancing pool by clicking Add. Each VM can be included to multiple balancing pools.
  
  In the Add members window that opens, select the desired VMs, and then click Add.
  
  You can select only between VMs that are connected to the chosen network.
4. In the Health monitor section, select the protocol that will be used for monitoring members availability:
  - HTTP/HTTPS. The HTTP/HTTPS method GET will be used to check for the response status code 200. Additionally, specify the URL path to the health monitor.
  - TCP/UDP. The health monitor will check the TCP/UDP connection on the backend port.
  - PING. The health monitor will check members’ IP addresses.
  By default, the health monitor removes a member from a balancing pool if it fails three consecutive health checks of five-second intervals. When a member returns to operation and responds successfully to three consecutive health checks, it is added to the pool again. You can manually set the health monitor parameters, such as the interval after which VM health is checked, the time after which the monitor times out, healthy and unhealthy thresholds. To change the default parameters, click Edit parameters, enter the desired values, and then click Save.
5. Click Create.
Add more balancing pools, as described above.
Click Create.