Can I use HA for VM/app monitoring only?

[vitaprimo]

I have this VM that I think it's been dying. I've found it off a few times but things have been cray so I'm not a 100 if I did it myself. While I find some time to diagnose the VM I'd like to enable HA just to keep it alive if it dies but I don't need any other feature of HA but to keep VMs–or a VM–alive. I'm not too crazy about keeping vCSA alive either, it makes too much noise on its own, too much logging (1000x+ w/Log Insight) but if it's need it by all means.

On the other hand vROps, which hasn't been set up completely (lack of time too) has already discovered a handful of services and has built VM relationships on its own all without a single agent installed, only a remote collector (one) and that's it:

If vSphere HA isn't the best fit for VM and/or app monitoring could Operations Manager fill that role? It's practically screaming it can deal but I've always thought it was more like a baseline-comprehension thing or an add-on you'd get because of feature-FOMO (totally guilty of that.)

Thanks -- BTW, if you have another idea/solution not involving these two, please share! (except for Tanzu, I can't justify the license.)

[daphnissov]

HA is only going to help you if a host dies. If your "cray" VM is on that host then, naturally, it'll be rebooted, but that's about what HA will do for you here.

[depping · Accepted solution]

Actually that is not true, HA also has "VM and Application monitoring" and it has the ability to monitor the VM (needs VMware Tools for this) and have the VM be "reset" (not rebooted) when the VM is nonresponsive. It looks at the "vmware tools heartbeat" and it will also look at "storage and network IO". With App Monitoring you can even monitor particular services if you want, this requires some more manual configuration though, and could be a bit more work, and to be fair, I don't know too many people using it as a result of the complexity.

So yes, you can use VM and App monitoring. 

[daphnissov]

I wrote my reply in haste, and yes of course you're right. I should have added that part so I'm glad you did.

[vitaprimo]

By not rebooting you mean, if it went down--as opposed of simply not responding--it'd leave as is?

The funny thing is that the last time I restarted the VM I stayed around poking around what was available long enough to catch that discard the redo log question/statement that I had been missing when I ran off immediately.  😅

I restored to the one snapshot I try to keep per VM and it was back to normal. I was kinda still interested in the vROps monitoring thing though, so I went ahead I tried installing the agent on VMs from the vROps UI and it failed every single time. I tried from the command line and failed again. I paid closer attention to the docs and I read some Java requirement. Hard pass. I'm deleting all snapshots, consolidating then taking new offline snapshots for all VMs (+auto backups via Synology's included utils) then vCSA and vROps are shutting down until the next template operation.

By accident I just found out last week vCSA itself runs on Java, until then I knew vSphere was a little pricey, a little painful for privateuse/nonprofit, though flashbacks of Hyper-V made it reasonable--then I found out it runs on Java. 😩 That's completely irrelevant though. ADHD, sorry.

Thanks a million to both for answering!  ⭐️

It'd be kinda cool if this had a since both answers are OK so here's a flip a coin button. Maybe more funny than cool…