Disappointment...what do you use vFoglight for?

Simon,

I'm sorry to hear you are having trouble with vFoglight.  I would be happy to setup a webex with you to go through the solution and discuss how my customers use the product, and also field any questions you may have.  Let me know how to contact you to setup a meeting request ok?

Thanks,

Tommy

Hi Tommy,

Thanks for the invite but I'd really like to hear directly from the users of vFoglight as to their real world experiences. I know the product looks great but what I need to know is what kind of problems has been helped to solve proactively or reactively, more effectively than the native management tools.

I login to the console every day but it cannot tell me where the issues are. RIght now I see > 100 alarms, and I cannot tell from vFoglight alone if any are valid. I have to rely on vCenter and traditional monitoring tools to spot real problems.

Simon

Can you please post these alerts?  A screen snapshot or export using excel?

Thanks

   Golan

Hi Simon,

I think you're right that vFoglight doesn't offer too much over the built-in tools in vCenter if you're not using the alerting component. I've had to really spend some time fine tuning the alerts for our environment. Initially we had a similar experience where it was just a massive avalanche of alerts. Any important ones we actually cared about got lost in the endless number of meaningless alerts. After fine tuning I can say that the alerts are really useful now.

In addition, the real value for us was a consistent monitoring environment. We have a very large virtual environment, but also have a lot of physical servers and specific applications we want to monitor. Being able to do that from a single tool was the big selling point for us. If the only thing you care about is monitoring the health of your VMware environment I think that's a judgement call for your team. If you can get the alerts under control I think you can find a lot of value. If you're already getting alerts from other tools, and find the monitoring tools in vCenter to be adequate then you guys may not have a good use case for it. Feel free to email me if you have any specific questions.

Ryan

Hi Ryan,

Thanks for the response, it's quite helpful. Indeed the alerting is the core problem for us - primarily we're facing unnecessary alerts about high memory or cpu utilization, or significant changes to memory and cpu utilization. The reality is that these alerts are due to scheduled events like backups, av scans or normal system activities such as vMotion (either manual or due to DRS). We have a medium sized environment >2000 servers in many locations so the idea of a single tool for all server monitoring is inviting but not realistic in practice.

We are looking forward to the next release which promises to solve the alerting issues with the product. Until then it will likely remain unused.

Regards,

Simon

Just one additional question - if we forget alerts for the moment - what else do you use vFoglight for? Can you give me a few example of how it helps your day to day work and save you time?

I'm really struggling here.

Thanks

The following is one perspective, from a guy spending a bit of every day learning about vFoglight.

We have an 11000 VM environment, with 10 Virtual Center servers.  We use vFoglight in a number of ways, based on the fact that it consolidates all of the VMware information in one place.  

• Alerting - we use a few, very specific alarms (Swapping, Agent data collection issues, ESX Hosts/VM extreme performance issues, etc.) for things which need to be looked at immediately, and only turn on new ones when we've thought through their use very carefully. 
  
  Out of the box, most of the alarms aren't meaningful, since they compare performance to fixed thresholds, which in many cases you don't really care about until they hit the Fatal level.  The alarms should compare alarm data to dynamic thresholds (Does this same VM spike utilization every Monday morning?  Not an alarm).  Eventually we'll replace those with customized performance trend reports/dashboards and alarms based on comparisons to dynamic bounds (or Quest will implement that feature in the product.)
  
  Note that you can change alerts to NOT send e-mails at lesser alert levels, and still have the alerts show up on the objects.  I have 250,000 e-mails in my inbox which say this is a great idea!  Also, remember that if you don't acknowledge alerts, they will generate follow-on alerts, and cause potential storage and performance issue in large organizations.  (Another good reason to not use alerts you don't really need.)
  
  
• Troubleshooting - we use vmExplorer as the first stop when troubleshooting issues.  Screens could be enhanced, but the data is there.
• Capacity Management - We use it as the primary method of deciding which cluster to put the next VM on.
• Capacity Planning - what new clusters will we need?
• License Management - both vRAM and guest OS licenses
• Chargeback - what did each VM use in resources for the month?
• Integration into other systems which need VC data - we replaced direct calls to the VC databases with vFoglight exports - much more reliable

One thing that's important to realize is that this is not a tool you can just install, turn it over to the users, and move on to the next project.  We made that mistake, and it took us a year to realize that we needed a different approach.

You don't get real value from the product until you start saying "How can I use it to centralize/reduce the number of other tools I have deployed/have to use in daily operations?"

vFoglight is incredibly powerful, and the fact that you can customize it to meet an amazing variety of different uses makes it unique in the market.  We've looked at every product on the market for the past 2 yrs, and there are products with certain features which are better, but nothing comes close in the overall package.

Sadly, finding out how to make many of the features truly perform isn't easy, since Quest's documentation and training is horrid.

John is right here on all counts.  I will add something to this as well, we just recently released new training videos for vFoglight 6.6 at the following training site:

http://svgtraining.quest.com/

Also I would point out that full documentation across both Foglight and vFoglight is contained here:

http://edocs.quest.com

I will admit that the documentation when you get into building out non-standard functionality can be lacking at times, but this community is here to help. 

I'm still pretty new to the product, but so far I'm happy with it. Our environment isn't as large as yours is Simon, just under 1000 VMs, 2 vCenters (one DR), but I'm finding vFoglight to be imensely helpful. Our other monitoring tool just doesn't have the ability to dig into VM specific issues as well as vFog can. I also have been working on modifying the alarms to make them useful, as others have said, the out of the box ones just aren't useful most of the time and give you way to many that you just end up clearing and moving on. Now that I have the alarms setup to be helpful it's become quite powerful for us. Then reporting was another big missing part for us and the reporting, especially if you utilize the vBundle and Velocity cartridges, has become something I use regularly when our management wants to know something about our environment that you just can't get out of vCenter easily. I think if you put some time into tweaking it to better fit your environment you may find it to be more useful than it seems to be today.

Thanks for the information John - interesting perspective. Perhaps we're not committed to investing time into configuring the tool in the way that you are.

I'm part of a very complex organization with multiple escalation groups and level - all using different monitoring tools, we are all fairly silo'd. Our team only really wanted an effective vm monitoring tool, for our own purposes, not for other teams. I think this is where you are benefiting from vFoglight and we probably cannot. We are also very focused on training our teams on the configuration, design and maintenance of our critical production (i.e. user facing) systems and don't have the same time and budget to invest in these non-critical supporting systems like vFoglight. 

I also wish the software was a little more intuitive. I've created a bunch of dashboards now, but the next day when I want to tweak them I cannot remember exactly how I got there. There has to be a better way!

But you can still benefit from it in your setup. Your environment sound a lot like mine, where I bought vFog just for my team and have no intention to share it with other areas unless I have to. The dashboards, reports, alarms, etc are only for us since our "Enterprise monitoring" solution doesn't do a good job on virtual servers. It does have a learning curve, but if you take the time to learn some of it you'll be surprised what you can do with it and it's not all that hard. Tommy from Quest is always willing to assist you with learning the product, he's helped me out several times with no problems, so reach out to him if you have an hour to learn part of the product and see what it can do for you.

Good luck, Jonathon