This weekend I figured I would give PHD Virtual's product called PHD Monitor version 10.1.3.1 a shot.
When I downloaded the trial, it's downloaded as a windows executable. I really wish it was a virtual appliance so I didn't have to have another 2008 R2 server in my environment to eventually patch. So I spun up a new 2008 R2 VM and figured I would give the installer a go without reading any documentation. The installation was very simple and straightforward. The installer includes a version of SQL Express 2005 as a database for the product. After the installation, the icon below is displayed on the desktop. I opened it up and began the configuration process.
The icon opens up Internet Explorer and navigates to http://IP_of_Server:81. I then went to my Firewall settings of my VM and turned off all firewalls so I know that my web service was available. When the webpage opens, it asks for your license key, in my case I kept EVAL. Then we begin with the configuration wizard.
I choose the first option, entered in my vCenter settings and it found my ESX hosts and all VMs without any issue.
Instead of working in a VM, I figured I would get back to my desktop on my mac and that's where I found a pretty big issue for all us mac users. You CANNOT access the interface with anything except Internet Explorer. No Firefox, No Safari, No Chrome, IE Only. Yikes.
Alright, back to my Windows 7 View session to continue investigating. The welcome dashboard shows that this program can monitor more than just your virtual environment. There's actually a crazy amount of stuff you can monitor. UPSs, Physical Workstations, Storage Arrays, Citrix XenDesktop Sessions, etc. Pretty much anything that can Syslog or send a SNMP trap. It's really a ton of stuff. Since I'm only interested in my virtual environment, that's what I will focus on. You can tell that PHD Virtual is focusing on the Citrix market because of the ability to dive deeper into Citrix based products.
After the initial configuration, here is the dashboard. This is a pretty nice high level for not having to deploy any agent to the VMs.
Let's dig into an ESXi host. You can see a simple overview of the host with CPU and memory stats. While looking at these stats, it defaults to a 12 hour view. You can look as close as 1 hour or as far back as 3 months. The program can't do real time, but you can refresh it every few minutes to update it.
As you scroll down you can see lists of datastores and VMs on that particular host and latency and IOPS stats for all the datastores. It would be nice to see a breakdown per VM right here, but we can atelast see all the datastores.
Now lets dig into the VM properties. Just like the ESX host, it's a familiar dashboard of CPU and Memory statistics.
As we scroll down, there are IOPS and disk latency metrics as well as network statistics. Those are some good metrics to see.
I figure I would give the Agent deploy process a go to see what you can monitor with a virtual machine. It's simple enough to discover, but the deployment process isn't as polished. You can only set credentials once for the domain and when it's time to deploy, there isn't a status bar or anything else that tells you what's going on. I often thought that the process just didn't complete or died off until a surprise window pops up that tells me the deployment is done. Afterwards, you can see if the deployment completed or not.
After completion I can go to the inventory and see that the agent successfully completed. In the dashboard, there is also this new icon next to the VM.
I'm not a fan of agents, but let's see what extra information is gained. Within the dashboard graphs, there isn't anything special except seeing the agent status which isn't important. The agent also give some more specifics around CPU and RAM in the monitor View.
The real goal of the agent is for doing service, application and file monitoring. It's a simple setup. Go to the Rules and Alerts under configuration, click new, and select the type of alert. I wanted to configure my View Security Server service. So I choose the WinProcess Service Option.
Clicked on my VM, found the service I wanted to monitor, checked it and clicked save.
After clicking save, I went to the View Security Server VM and shutdown the service. After about 30 seconds, the light on the dashboard changed from green to yellow and I investigated the warning to see it had been restarted. Sure enough, I went back into the View Security VM and did a F5 refresh to see that the service did in fact restart. Pretty cool.
The product has a simple dashboard that allows for some simplistic monitoring based on thresholds. Even though I'm not a big fan of agents, the process watch list piece is helpful in automatically restarting services. One thing that seperates this product among other monitoring tools is that you can monitor anything with SNMP or Syslog properties or any physical servers via agents. Here is a picture of my DS411 NAS and the network properties.
Give the product a go and see what you think. It comes with a 30 day free trial which is plenty of time to gather stats and get your whole environment, VMs, physical servers, routers, firewalls, SANs, and more pulled together into a single monitoring solution.