Hi all,
Can anybody explain what how does vcops manager collector helps in reducing bandwidth constraints. Also from where can I download the exe for this which can be installed on remote windows/Linux servers.
Thanks in advance.
The collectors really come in to play when. 1. You've got a large environment of scale and a single collector on the default vApp needs to be split off or distributed due to load on the solution. 2. You've got firewall restraints and can't communicate directly from vCOps to data source. Plus the security risk you get by poking a hole in the vC Ops firewall. As a result, it's not a popular option and is only done is special circumstances.
For network bandwidth requirements, it's going to depend on what adapter you're using for the data collection. Some adapter query/retrieve more data than is actually used/imported by vC Ops due to filters/whitelists/configurations. Some adapters also have configurations of quantity of threads and size of requests to trim/throttle data collection. The answer is, "it depends". Like is many situations, it depends on your environment size, latencies, and configuration.
When you install vCOPS it comes with a default collector.
When you have a distributed load the you will be installing a collector instance in that particular DC and collecting the data locally.
This Remote collector will talk to the centralized vCOPS server and send the metrics data.
For details on collector check this link
The exe file is present on the VMware download site
See the link below for download
For details on how to configure it see the below doc
The remote collector isn't something that is widely discussed, as it is ill-recommended for anything other than larger environments, standalone installations or situations where firewall concerns give no alternative. Configuring the remote collector also opens the appliance's firewall and creates a security risk, which further adds to the special circumstances required before deploying an additional collector.
Mark, Just curious.
The remote collectors too use the same bandwidth to send all the metrics across the network right? How is it useful? If we had to measure or recommend a network bandwith between the remote site and central site for metrics sync, how woould we measure it?
The collectors really come in to play when. 1. You've got a large environment of scale and a single collector on the default vApp needs to be split off or distributed due to load on the solution. 2. You've got firewall restraints and can't communicate directly from vCOps to data source. Plus the security risk you get by poking a hole in the vC Ops firewall. As a result, it's not a popular option and is only done is special circumstances.
For network bandwidth requirements, it's going to depend on what adapter you're using for the data collection. Some adapter query/retrieve more data than is actually used/imported by vC Ops due to filters/whitelists/configurations. Some adapters also have configurations of quantity of threads and size of requests to trim/throttle data collection. The answer is, "it depends". Like is many situations, it depends on your environment size, latencies, and configuration.
That's exactly what i was thinking too. Thanks a lot for clarifying it
So if I'm understanding this correctly, the collectors will do some of the data aggregation and perhaps remove some of the strain from your vApp's Analytics VM?
The remote collector will usually be used for specific adapter needs (windows auth, .net services, etc), network needs, or otherwise. Data aggregation and removing load from the primary vApp is usually not the reason for spinning up another collector (although applicable to very large deployments with custom/3rd party adapters). If there is a bottleneck it's going to usually be on the Analytics processes/underlying virtual resources (disk I/O, etc). If there is a takeaway from this conversation, it should be that deploying a remote collector on the vApp deployment of vC Ops is not in the cards unless you meet a very specific criteria.
Hi everyone,
Thanks alot for this info. In our environment, we have 3 zones for vdi spread across different regions. As of now Vcops for View has been installed in one of the regions. What could be the efficient way of deployment in this case. Can three different Vcops managers be deployed in all the three zones? How does the licensing works in that case?
HI Everyone,
I too have query regarding monitoring of geographical distributed VCENter . In my case we have 03 VCenter located in (brussels , belgium, china ) , i need to connect all the Vcenter to VCOPS manager installed in sweden datacenter . All the monitoring filters are default no any customized filters are created and the VCenters have approx 200 VMs & 10 ESXi hosts . Could you please suggest how much approx bandwidth consumption will be there .
Let me know if more info is required .
What attribute packages / profile are you using? All attrib pkg / default attrib package OR Balanced/Full profile?