Cloud Monitoring
In order to ensure the smooth running of business operations in cloud and virtualized environments, network management must move away from point monitoring of IT infrastructure to instead monitoring business service availability and performance. Zyrion’s Traverse solution enables network management to go beyond just looking at the performance of individual nodes or components by also providing integrated cloud monitoring and Business Service Monitoring (BSM) capability that offers a holistic service-oriented view of the performance of the IT and cloud computing infrastructure.
CloudStack OpenStack Monitoring
Zyrion can monitor public and private cloud infrastructure including Cisco UCS, VMware, Xen, RedHat KVM, HyperV, CloudStack, Openstack, EC2 and monitor their performance in a unified and integrated manner. No agents are required, since Traverse uses their native XML or Java APIs to monitor their performance.
Read Whitepaper on Traverse Support for Virtualization and Cloud Monitoring Environments »
Traverse integrates seamlessly with virtualization technology, such as VMWare vSphere and vCenter to enable end-to-end monitoring of physical and virtualized environments. Examples of the types of VMWare metrics, and the purpose of monitoring these is summarized below:
cpu.usage.average (%)
Measures CPU use at the VM level. Utilization can be compared against thresholds to provide warnings for VMs that may be having performance problems due to over-utilization.
cpu.ready.summation (millisecond)
Indicates whether VMs are waiting for CPUs to be allocated to proceed forward with processing. This metric points to potential bottlenecks for available CPU resources.
cpu.swapwait.summation (millisecond)
Indicates VMs waiting for swap space, and thus CPUs not being effectively utilized due to lack of sufficient memory.
mem.swapin.average (KB)
Provides indication of the rate at which memory is being swapped in from disk. A large number represents a problem with lack of memory and an indication that performance is suffering as a result.
mem.vmmemctl.average (KB)
Indicates amount of memory currently claimed by the balloon driver. A value in this metric indicates that a memory bottleneck may be about to occur as this metric represents a VM taking memory away from other less needful VMs. Looking at swap rates would confirm if indeed there is a potential problem.
mem.overhead.average (KB)
This the amount of memory used by the VMkernel to manage allocated memory to VMs. Used to help determine memory allocation across VMs, since the overhead is greater the larger the amount of memory allocated to a VM.
disk.busResets.summation (#)
Indicates a bottleneck exists in the environment, given that disk reset cleans out all the commands that have been queued up in a disk bus. The problem could be disk overload or hardware failure.
Disk.commandsAborted.summation (#)
Specifies the number of aborts that have occurred in the last window of time. A non-zero value indicates a severe issue with the disk.
net.received.average (kbps)
Provides indication on the rate of data traffic being received by a VM.
net.packetsRx.summation (#)
The number of packets received by each vNIC (virtual network interface controller) on the VM during the collection interval.
Cloud Monitoring with Service Containers
With the transition of physical servers to cloud environments, the number of virtual servers has increased exponentially. What was once easy to visualize and track is now harder - trying to figure out that the application database VM lives on physical server A which also hosts the Exchange SQL server on another VM and the performance of the entire service is impacted by server A has become very difficult.
Using Zyrion's leading Service Container technology, administrators can see the impact of the physical and virtual infrastructure on their applications. Applying service container technology to Cloud Monitoring and Management should become an essential component of your monitoring software.
| 
