Open up the capacity management pack for other systems/counters without VMM
Would be great to be able to do some capacity planning on other counters or systems (sql dabatabase, scsm, scom orchestrator). You have a great engine in the backend so please utilize it fully. for example: would be great to be able to import other SCOM data like SQL perf counters so I can do capacity planning on my SQL servers. Or to be able to do capacity planning on physical hosts that are not managed by VMM.
Documentation for the updated capacity solution is now available:
This solution does require Hyper-V and does not work for VMware or Azure VMs.
Balasubramanian Murugesan commented
Any Update on Capacity and Performance Management solution availability under Solution Gallery for OMS
Can we get an update on the progress or a general time line for planned delivery?
It would be useful if there was a solution that did capicity planning for on premise servers
Reng Kwan commented
Is the capacity management unavailable at the moment? I don't see it under my solution gallery. If yes, when will it be back again?
Will the new OMS Capacity planning solution work with VMware VMs and physical servers or are you going to require Hyper-V? We would like the capacity planning solution to not have a visualization requirement and support any VM or physical server reporting to the OMS cloud service.
We want VMware support with Capacity planning. Can you please make this a priority? Can you provide an update on the direction you are heading with capacity planning? Is there any hope that you will soon support VMware? Up voting.
John Pritchard commented
To have the same Capacity solution for Azure VMs as we do for on prem VMs (which depend on VMM and SCOM).
Use Direct Agent and extend to include VMM capabilities to do capacity data collection.
Solution would display the same as the Capacity Management for on prem VMs
Michael Voegtline commented
Need for VMware!!!!!
Manjunath. C commented
Capacity management is not reporting the data after adding the solution and installing the monitoring agent. Let me know is there anything else needs to be done. Added machines are in azure vm.
What about capacity management of a public cloud based on subscription limits. It would help to provide information of current and historical growth of on-premise, provider using WAP and public cloud for example Azure. This is the model that will come before the complete hosted and will help the customer to see the growth rates all up.
Also for one other feedback that was merged with this about 'drilling into VMs' - that is the same as this other idea http://feedback.azure.com/forums/267889-azure-operational-insights/suggestions/6519214-capacity-planning-efficiency-chart-doesn-t-reveal
But essentially, all of these fall in the same category: support collection of performance counters directly from agents (including direct agents, on VMs or on Hots) - which is something we see being part of custom log collection.
Doing 'capacity' analysis or 'monitoring', on top of that data, is a different set of queries and visualizations, but we need to provide the data source first.
That is the extent that is probably possible with just standalone agents.
For the *current* capacity implementation - that does MORE than just collecting performance counters; it takes into account how configuration has changed, machines have been moved around failover clusters, have been turned on and off... but all of that knowledge about the environment is currently provided by the VMM integration and built specifically around that 'private cloud' or 'hoster' scenario, i.e. used in conjunction with the Cloud Platform System http://www.microsoft.com/en-us/server-cloud/products/cloud-platform-system/
Capacity for other scenarios (including 'just a bunch of hosts') require defining how the system looks like and won't just automatically light up in pretty predictions as you need to build a model first (i.e. SQL server capacity is different than 'single hyperv host' capacity and is yet different than 'cloud storage capacity', and so forth) to understand growth pattern and what does it mean to be 'over' or 'under' allocating, etc... different workloads and systems have very different angles to this.
But we do want to enable collection of performance metrics to allow building that intelligence on top of the data in a generic way.
You lost money if you offer SCOM integration and not integrate physical hosts. This missing part is one of the main reason why not investigate more like tool compare. I cannot present this nice tool idea and tell the management in the same time a second solution is required...