The scaling up of the proxy node is independent of the platform nodes in the cluster. Typically, the users install one or more proxy VMs per site. Within a site, the number of proxy VMs needed is a simple function of total number of VMs for which it has to collect data. Refer to the capacity of proxy VMs in the brick size table in the System Requirements section.
You can add a data source (maybe a vCenter or a switch) to exactly one proxy VM.
Scaling up Scenarios for the Proxy Cluster
Scenario1: Suppose there are 2000 VMs in a vCenter.
Install one medium proxy VM. Assign the vCenter to this proxy using the product UI.
Scenario 2: 1000 VMs in vCenter1 and 2000 VMs in vCenter2 (all of them are in one data center)
Install one Medium Proxy VM. Assign both vCenters to this proxy using the product UI.
Scenario 3: 9000 VMs in vCenter 1 and 2000 VMs in vCenter2 (all of them are in the same data center)
Install one Small Proxy VM. Assign both vCenters to this proxy using the product UI
Scenario 4: 1000 VMs in vCenter1 (data center1) and 2000 VMs in vCenter2 (data center2)
Install one Medium Proxy VM in each data center. Assign vCenter1 to proxy VM in same data center using Product UI. Assign vCenter2 to Proxy VM in its data center using the product UI.
Scenario 5: 9,000 VMs in vCenter1 without flows (data center1)
Install one large proxy brick. Assign this vCenter to this proxy using the product UI.
Scenario 6: 11,000 VMs in vCenter1 with flows (data center1)
This scenario is not supported. Maximum number of VMs that can be managed by one proxy VM is 10,000 without flows OR 6,000 with flows. And one vCenter can be added to only one proxy at a time.
Scenario 7: vCenter1 with 2000 VMs in January, vCenter2 with 5000 VMs in June
Install one Medium Proxy VM in January and assign vCenter1 to it. Install the second large proxy VM in June and assign vCenter2 to it.
Proxy VMs with a Platform Cluster
The number of proxy VMs does not depend on the number of VMs in a platform cluster. All proxy VMs communicate only to the first platform VM (
platform1 in the following example) in a platform cluster. A few example deployment models that are supported are as follows:
Case 1: A proxy VM connects to a platform cluster.
The proxy connects to
Case 2: Many Proxy VMs connect to a platform cluster
All the proxies are connected to
platform1. And then
platform1VM load balances both proxy requests and the data processing to other platform VMs in this cluster internally automatically.
Case 3: A proxy VM connects to the single platform node deployment
Case 4: Many proxy VMs connect to one platform node deployment