Overview
A load balancer is mandatory for the multi-node HA-ready production setup.
Automation Suite supports two types of configuration for the load balancer, as shown in the following sections.
You can configure the load balancer to use session persistence or sticky sessions, but it is not a requirement.
Note:
Currently, only Layer 4 (network layer) load balancer is supported.
Server and Node Pool configuration
This is the recommended configuration for the load balancer.
Configuring the backend pool
You need to create two backend pools that meet the following requirements:
-
Server Pool
- Consists of all the server nodes.
- There must not be any agent nodes in the Server Pool.
-
Node Pool
- Consists of all the server nodes and agent nodes.
Enabling ports on the load balancer
Ensure you have the following ports enabled on your firewall for the source of load balancer:
Port | Protocol | Purpose | Traffic forwarding |
|---|---|---|---|
| TCP | For HTTPS (accessing Automation Suite) | Traffic on this port should be forwarded to Node Pool. |
| TCP | For accessing Kube API using HTTPS; required for node joining. It is recommended to close this port after installation. | Traffic on this port should be forwarded to Server Pool. |
| TCP | For accessing Kube API using HTTP; required for node joining. It is recommended to close this port after installation. | Traffic on this port should be forwarded to Server Pool. |
Note:
For any ports other than HTTPS, we recommend it is not exposed outside the cluster. Run your nodes behind a firewall / security group.
If you have a firewall setup in the network, make sure that it has these ports open to allow traffic from those ports.
Refer to the following illustration for more details on the configuration.
Configuring the health probes
Probe | Protocol | Port | Interval | Re-entry threshold | Associate pool |
|---|---|---|---|---|---|
| TCP |
| 15 sec | 2 | Node Pool |
| TCP |
| 15 sec | 2 | Server Pool |
| TCP |
| 15 sec | 2 | Server Pool |
Refer to the following illustration for more details on the configuration
Alternative configuration
Important!
This configuration does not have resilience to nodes going down during installation.
If the primary server is down or deleted, cluster configuration needs to be updated.
FQDN of the primary server needs to be remapped to a different machine in the cluster that is available.
Configuring the backend pool
Create one backend Pool, as follows:
- Create a Node Pool
Enabling ports on the load balancer
Ensure that you have the following ports enabled on your firewall for the source of load balancer:
Port | Protocol | Purpose | Traffic forwarding |
|---|---|---|---|
| TCP | For HTTPS (accessing Automation Suite). | Traffic on this port should be forwarded to the Node Pool. |
Configuring the health probes
Probe | Protocol | Port | Interval | Re-entry threshold | Associate pool |
|---|---|---|---|---|---|
| TCP |
| 15 sec | 2 | Node Pool |
Refer to the following illustration for more details on the configuration.
Updated about a month ago