Crawl, walk, run approach to enabling VMTurbo Automation

Document created by justin.kutticherry Expert on Aug 31, 2015Last modified by fran.schwarzmann on Aug 15, 2016
Version 3Show Document
  • View in full screen mode

We’ve been working with a number of existing customers helping them get more out of their investment in VMTurbo and a big part of our efforts is around helping them reduce the amount of risk in the environment and eliminate performance problems from happening in the first place. After a few sessions of reviewing the recommendations within VMTurbo, the conversation always heads toward automation. Over 70% of VMTurbo users choose to automate within the first 30 days! What I've found is that when we begin the journey toward automation in VMTurbo, we always start in the same place.

 

Pick a cluster that has a significant amount of workload running on it and one that has performance issues. You can easily see this by scoping to a particular cluster in the ‘Optimize Summary’ view. In the example below, you can see hosts experiencing ballooning, ready queue and swapping – all of which effect application performance.

 

 

Start by reviewing the actions that VMTurbo is recommending and drill through the data to make sure you understand why VMTurbo is suggesting the recommendation. Does the recommendations make sense and are they actions you would take on your own? You can look at the data supporting the decisions by clicking on the ‘Show Details’ icon on the right to understand why VMTurbo is making the recommendation.

 

Once you feel comfortable, start executing the actions manually within VMTurbo. Repeat this process for a week and continue to take actions manually to build trust that the VMTurbo recommendations are accurate and are not causing harm to the environment. Monitor the cluster and if there are no issues, enable automation on that same cluster.

 

However, before enabling automation, make sure workload placement policies are imported from vcenter properly and if needed, create additional policies within VMTurbo that respect all your business constraints. Policy > Workload Placement Policy

 

Once the right polices are in place, enable automation for placement decisions. Policy > Action > VM >

 

 

Take a snapshot of the environment before enabling automation and then wait. Allow VMTurbo to control the environment for a week and keep an eye on the actions that are being executed by VMTurbo. Optimize Summary View > Actions Executed Log will contain a detail log of all the executed actions. This log can be exported and reviewed offline:

 

 

Take a snapshot of the cluster after a week. Do the performance issues still exist?

 

Try automating a few more clusters (placement decisions only) to continuing seeing the impact of VMTurbo over time. Next move to automating storage moves on a cluster by cluster basis and continue to automate more and more actions from there.

 

 

Don't forget to share your progress with the rest of the Green Circle Community Share your Automation - Get 1000 points & profile badge!

 

CSC Green Circle Forum

Re: Dipping the toes in the pool, starting automation slowly....

Attachments

    Outcomes