• When any VM has a CPU Ready time of over 20% and the host CPU utilization is also over 90%, it signals CPU overcommitment & would need addition of CPU or VMotion enabling
• More than 1MB/s of swap in or swap out rate signals memory overcommitment
• From an EMA benchmark report, average performing enterprises had physical CPU utilization of 45% while best performing enterprises had this at 70%. Where do you stand?

The above were just a few “rule of thumb” points and VMware metric descriptions that were covered in our webinar titled “VMware Performance Monitoring – the Must Haves”. The presentation that was used for the webinar is embedded below. While we get ready to host the recorded version, you can write to us if you are interested in the transcript of the webinar: opmanager-marketing[at]manageengine[dot]com

Network mapping: one of the essential needs for any network administrator, whether to identify where the problem is or for a quick visual inventory to identify what/where the device is.. Though an admin can get the picture from static network maps,

• A static map fails to scale with the growing network, especially in the SMBs, where business demands change quickly
• It is also ineffective for an administrator to refer to a static map whenever a fault happens in the network, one needs live maps!
  

We spoke to a handful of customers to understand their needs on network mapping and here are some of the requirements you can also easily identify with:

• A quick means to come with a complete network diagram
• Scheduling of network scanning to include newly added devices in the network
  
• Ability to customize these maps (e.g. adding a background image)
• A live network map, showing links or devices having trouble
• And finally the reporting, for a particular network segment or a group of devices

When we came up with the latest release that had automatic network mapping, we thought an option to export a network map to the business views would also help.

OpManager now has options to export automatic network maps to business views (custom map).



You can add custom background images to your exported network maps in a Business view.



In addition to all these, you can even export the business views or the network maps directly to Visio.



Next Steps:

• If you have not tried it out, refer to our user guide on how to do it : http://www.manageengine.com/network-monitoring/help/userguide/layer2_mapping.html
  
• If you have already tried it, please share your feedback to OpManager team at "opmanager-marketing@manageengine.com"
• If you need any further enhancement, feel free to post them at OpManager ideas here "http://forums.manageengine.com/opmanager/type/ideas "

Hope you like the way it has come-out!

Happy networking,
Kalvin Ram
Team OpManager
Network monitoring software from ManageEngine

It's the weekend that hosts the best football played in the past four years - Germany takes on Uruguay for 3rd play-offs and Spain fights Netherlands for the grand World Football Champions title. Octopus Paul who's been on a 100% success rate in his predictions so far at FIFA 2010, says that Germany and Spain win in their respective games. And having seen these four teams play I too think it's Spain-Netherlands-Germany-Uruguay placed 1-2-3-4. So Paul will remain 100% successful in his predictions this worldcup?!!

What if you had magical, psychic Octopus Paul solving IT failures for you? Check out what we think at http://www.manageengine.com/network-monitoring/ .. After the football weekend, after you've cooled off, do the latest OpManager 8.7 and make yourself magical & psychic with IT!

Get the beers & chips in stock and have a great football weekend folks!



 

We get frequent queries from our customers on their applications being slow and ways to monitor this. Monitoring basic CPU, Memory and Disk utilization gives a fair indication of where the problem could be, but you need to monitor more than just the resource utilization to actually determine the reason for application slowness.  Multiple system resource parameters or even a single parameter can impact the performance of an application. Let's look at "Disk I/O" - one of key parameters that causes application performance related problems.

Why monitor Disk I/O:

The hard disk is one of the slowest components on a server. The hard disk drive bandwidth is around 300 times slower than that of the memory and added to that, memory latency is very low - if you do the math, memory performs around 2000 times faster than that of the hard disk drive. Check out http://nathanaeljones.com/153/performance-killer-disk-io/ for a neat explanation.

So monitoring and keeping the Disk I/O at healthy levels can prevent a lot of application performance hiccups.

Does my server have Disk I/O problems?

Depending on your server, there are different ways of finding Disk I/O. Refer to http://www.performancewiki.com/diskio-tuning.html for the Disk I/O commands on various servers.

Another simple way, and what we recommend ;), is to use OpManager. OpManager monitors Disk I/O out-of-the-box for Windows servers (see screenshot below) and with a few customizations also for Unix-based devices.




Before I sign off, thanks to Brien M. Posey at SearchDataCenter for selecting OpManager for 'enterprise-class server monitoring' in his latest write-up, 'Selecting benchmarking tools for specific needs'.

For more on OpManager's related features refer to server monitoring with OpManager.

--
Pravin
Team OpManager - The Network Monitoring Software from ManageEngine

Have you ever felt that...

• the alert on a resource crunch was a tad too late?
• the VMware crash could have been averted if the hardware resource allocation had been properly planned?
• the time to get down to the root of the problem proved costly on your SLAs?
• correlating information from multiple solutions to assess performance problems is bothersome?
  

If you answered yes to one or all of these, you must be joining our upcoming webinar on how you can have just one solution identifying all of these performance problems and resolving them quickly.

Register for the webinar
We had an overwhelming response for our previous webinar (thank you!) on network performance management where we had folks joining from across the globe, spanning different time zones. We are hosting the webinar on two consequetive days this time, so that you can choose the session closest to your time zone.

Topic: Top 5 server performance problems and how to resolve them.

Schedule: Register for any one of the following sessions:

• Tuesday June 22, 2010  9:00 AM - 10:00 AM GMT: Click here to register.
• Wednesday June 23, 2010 10:00 AM - 11:00 AM EDT:  Click here to register.

Michael Dolgae of First Place Bank was one of the customers who visited our booth at Interop. We discussed at length on his network requirements and he had some good ideas we could use. Here’s Michael sharing his story of a Solarwinds Orion consultant coming over to inspect his network and going back all impressed with OpManager!

Michael also impressed us in the way he so comfortably used his brand new iPad (like he was using the gadget all his life!). He was also a heavy user of OpManager’s smartphone interface, of course he used the iPhone. Thanks Michael, for the good ideas you shared and for the great feedback. With what we have in store, we look forward to surprising you and any more consultants you may have visiting :)

        The breadth and depth of OpManager's fault management capability is fully utilized by many of our users. Few users however may not be aware of the extent of flexibility the application provides as far as managing alerts are concerned.

Yet another frequently asked/discussed requirement is the ability to forward alarms in OpManager as traps to another trap destination. Requests are also raised to know if the incoming trap can be redirected to another application. YES, both are possible. Here is how you go about the forwarding:

Forwarding Traps

Here, OpManager simply forwards the SNMP Trap to another manager/device. Forwarding traps from OpManager involves a couple of configuration changes in two files. Refer the table below:

1. Effect the following configuration changes:

File name & path	Existing entry	Modified entry
/opmanager/conf/trap.forwarder		(replace with the correct IP address and port)
/opmanager/conf/trap.filters

2. Save the file and restart OpManager.

OpManager starts forwarding the incoming traps to the destination IP address and the Port mentioned in the trap.forwarder file.

Forwarding OpManager Alarms as Traps

You might want to forward the alarms in OpManager as traps to another destination/trap receiver. Here is how you do it:

1)Ensure Net-SNMP is installed on the OpManager server or download from the links given at the end of this post:

2)Create a batch file sendtrap.bat to invoke the snmptrap command. Place this file in  /opmanager folder. The syntax for this command is as follows:

snmptrap options hostname community trap parameters...

For Version 1, the required trap parameters are:

enterprise-oid agent trap-type specific-type uptime objectID type value...

The code snippet for this batch file is given below:

snmptrap -v 1 -c public %1 .1.3.6.7.2.1.3.1 %2 1 6 10000 .1.3.6.7.2.1.3.1.1.2 s "%3,%4,%5,%6"

where

-v 1	The SNMP version
-c	The community string and the default SNMP community string is public
%1	The manager name, viz., the hostname or the IP address of the manager that must receive traps.
.1.3.6.7.2.1.3.1	Specify the appropriate enterprise OID
%2	The agent-name, that is the hostname or the IP address of the device for which an alarm is generated.
.1.3.6.7.2.1.3.1.1.2	Replace with the appropriate specific OID
1	The generic type number
6	The specific type number
10000	The sysUptime in milli-seconds
s	Refers to the data type of the OID (string type)
"%3,%4,%5,%6" (replace the % symbol with $ if you are doing it on a Linux box)	The message from OpManager that can contain one or more of the following variables: $displayName - the Displayname of the device $strModTime - the time when the alarm was generated $category - the category to which this device belongs to $severity - the severity of the alarm $message - the message of the alarm ( like Web Service Down or Threshold Violated ).

3)Configure a 'Run Program' notification profile specifying the following:

Command Name: sendtrap.bat

Program Arguments :[manager-name] $source $displayName $strModTime $category $severity $message

3. You can associate this notification profile to devices using the Quick Configuration Wizard. Whenever an alarm is generated on the managed devices, a trap is sent to the other monitoring server.

NET-SNMP download (zip)

NET-SNMP download (exe)

When one tries picturizing a NOC, the big screen monitors immediately come to mind. Today’s post is about those big screens and how they can be made more useful with OpManager.

About NOC screens:

Wall mounted LCD/ Plasma screens are typically used by different IT teams for various network monitoring, fault management, and helpdesk purposes. The screen's visuals are rotated every few minutes at a pre-configured interval. For example, a view of

• Essential devices' availability statistics
• Top interfaces with high bandwidth utilization
• Availability status of the business critical services and applications
• Recent alarms or un-acknowledged events
• A network map showing the remote branches or datacenters connectivity and link performance
• Helpdesk ticket queues
• Recent call tickets & more…

Sometimes the IT team is even hooked on the screen by live TV feeds of baseball games or weather maps!

Here's how OpManager can help make best use of your NOC monitor screens:

To begin with, one can customize the IT dashboard to display required network performance information using the 90+ out-of-the-box widgets. The widgets always get the most recent polled information. One can then create multiple dashboards e.g. an exclusive availability and performance dashboard for the network devices, a separate one for the datacenter and server etc.

Obviously the big screens have an expanse of usable real-estate. One can increase the number of columns on each dashboard and decide on how much space each column should occupy.



After creating a couple of dashboards, you can group these dashboards in a CCTV View with a custom-configured refresh interval. You can also have multiple CCTV views. You can then click on these CCTV views to pop them out in a new browser window to place them on to another big screen.

One can also get another web application or other web feeds within OpManager’s dashboard using the “Custom HTML or Text” widget.




By the way, these features are available in all the stand-alone OpManager editions and requires no extra module or web-servers. Try out these nice dashboards and have a blast in the NOC (certainly no pun intended)!

Some useful links for you:
1.    Creating an individual dashboard and adding the widgets you’d like to see in it
2.    Creating a CCTV view by grouping these individual dashboards with the desired refresh interval

--
Kalvin
Team OpManager
Network monitoring software from ManageEngine