vRealize Automation Load Balancer Settings

I found some conflicting information about setting up load balancers for vRealize Automation in a Distributed installation, specifically around Health Checks. The following health checks were found to work for a fully distributed installation of vRA 6.2.2.

vRealize Automation Appliances

This is the pair of vRealize Automation Linux appliances that are deployed via OVA file.

Type: HTTPS

Interval: 5 seconds

Timeout: 9 seconds

Send String: GET /vcac/services/api/statusrn

Load Balancing Method: Round Robin

vRA Node 1 port 443

vRA Node 2 port 443

 vRealize Automation IAAS Servers Web Components

These are the load balancer settings used on the Web components. This would not be necessary for the DEM Agents and DEM Workers.

 Type: HTTPS

Interval: 5 seconds

Timeout: 9 seconds

Send String: GET /WAPI/api/statusrn

IaaS Server 1 port 443

Iaas Server 2 port 443

vRealize Automation IAAS Model Managers

These are the load balancer settings used on the Model Manager component. This would not be necessary for the DEM Agents and DEM Workers.

Type: HTTPS

Interval: 5 seconds

Timeout: 9 seconds

Send String: GET /VMPSProvisionrn

Receive String: name=”ProvisionService”

IaaS Server 1 port 443

Iaas Server 2 port 443

vRealize Orchestrator

This is for a vRealize Orchestrator Cluster being deployed from the VMware OVA.

Type: HTTPS

Interval: 3 seconds

Timeout: 9 seconds

Send String: GET /vco/api/docs/index.html HTTP/1.1rnHost:rnnConnection: closernrn

Receive String: 200 OK

Load Balancing Method : Round Robin

vRO Node 1 Service port 8281

vRO Node 2 Service port 8281

 vRealize Automation vPostgres Database

This assumes that you’ve setup a pair of vRA appliances and they are replicating the vPostgres database between the two instances.

NOTE: in order for the send and receive strings to be valid, a script must be run on each of the vRA appliances after vPostgres replication has been completed. This script allows for the receive string to be visible to the load balancer. Without this it won’t work. VMware KB Article: 

Type: HTTPS

Interval: 5 seconds

Timeout: 9 seconds

Send String: GET /vPostgresService.pyrn

Receive String: Postgres.Master=true

Username: <VAMI Username>  {the load balancer uses the root username and password to check the receive string}

Password: <VAMI password>

Alias Service Port: 5480 {the load balancer check the receive string on a different port from where the service runs}

vRA Node 1 port 5432

vRA Node 2 port 5432