vRealize Operations - Automated Remediation?

[btrainor]

Does anyone know if true automated remediation exists in vROps? The marketing material on VMware's website states this is one of the new features (vRealize Operations Features: Policy-Based Automation), but the only remediation functionality native to vROps that I can find is *manually* launching an out-of-box action in response to an alert.  Am I misunderstanding the concept of automation in this context?  My vision of automated remediation is having an alert automatically trigger an action based on a preset policy.  Any thoughts or guidance on this would be greatly appreciated.

Thanks,

Brian

[funksoul · Accepted solution]

Hi, Brian.

Yes vR Ops 6 has some builtin actions out of the box but they are manual actions and do not provide automation.

To build an automated infrastructure, you should integrate vR Ops to an existing automation framework such as vCO.

There is a 'vCO Remediation Workflow Package for vR Ops' that utilizes vR Ops SNMP Trap Plug-In.

(Cloud Management Marketplace | Solution Exchange)

Or you can use a 'REST Plug-In for vRealize Operations Manager Outbound Alerts' if you own a REST-enabled automation framework.

(http://pubs.vmware.com/vrealizeoperationsmanager-6/topic/com.vmware.ICbase/PDF/vrealize-operations-m... / p.120)

Regards,

Ho-Sung.

[btrainor]

Ho Sung, thank you for your response and clarification.  I thought that might be the case, but just wanted to get some confirmation from other folks in the community.

I’m actually familiar with the vCO Remediation Workflow Package for vC Ops and in fact, presented on that very topic at a TechTalk session at VMworld last August.  But because this solution this is not a true out-of-the-box solution, I was hoping that VMware would incorporate the logic into a later version of the vR Ops product.  In the marketing around the latest release I saw phrases like, “policy based automation,” “automated remediation,” “trigger actions,” and “automatically initiate corrective measures at critical thresholds.”  I was excited that VMware actually discovered a way to provide this functionality in a safe and reliable way.  Or so I thought…

I realize there are inherent risks to automatic *self* remediation so I wonder if VMware is holding off to work out these details.  I know that VMTurbo have built in automated remediation functionality so I’ll have to investigate how they manage to perform it in a non-risky way.

Thanks again,

Brian

[funksoul]

Hi, Brian.

Sorry I did not know that you're such an expert to this subject :smileyblush: [:blush:]

I was impressed at the new policy system that incorporates symptoms and alerts in the vR Ops 6 and I felt that it's the perfect solution for the policy part of "policy based automation".

But for the automation part, I was disappointed a bit because oob python action is manual, limited and it looked difficult to customize.

So as of now, I think vCO integration will be the best choice for these reasons:

- has huge extensibility that covers different domains

- all the automated actions are triggered from the alerts in the vR Ops

- it is an out-of-the-vmware-box solution

For the risks of automated remediation, what about integrating remediation procedure with organization's approval process?

Such as vR Ops(fire an alert) -> REST(remediation approval/audit) -> vCO(trigger an automated workflow)

Regards,

Ho-Sung

[btrainor]

Ho-Sung,

This is a long overdue reply, but thanks for your input. First of all, I would not call myself an expert at this stuff.  I've just been tinkering with it for awhile because I'm fascinated with this idea of the "self healing data center."  I would love to for us to have the ability to automate more of our fixes so we can focus on more proactive measures.  vROps does get us closer with guided remediation and smart alerts, which mean shorter time to resolution. I do like the gains made there.

So if we want true automated remediation using a VMware solution, I agree that vCO integration is the only way to go about it. This just requires an additional understanding of vCO workflows and some scripting which increases the learning curve for some users. But then again, it also gives us more flexibility. I just expected vROps to offer a simpler, truly out-of-the-box approach.

Your idea for mitigating risks by using a REST based solution with change approval sounds like a valid approach:

>> vR Ops(fire an alert) -> REST(remediation approval/audit) -> vCO(trigger an automated workflow)

This of course adds in the element of manual intervention, but at least it preserves control where strict change controls are required.  The automation would still be kicked off by a simple "OK" response to the alert.  Not a bad approach and would be a fun project to pursue!

Thanks again for the thoughts!

Brian

[pierrelxlab]

Hello,

If you want to do automated remediation, I added the capability to VROPS to call VRO workflows when an alarm triggers, this is done by installing a new REST notification plugin.

The information are here: New VROPS6 Notification plugin: VROPS6 alarm => Custom REST Request

Regards

[btrainor]

Excellent information! Thanks Pierre! Great work!