Вы находитесь на странице: 1из 8

WHITEPAPER

The difference between in-depth analysis of virtual


infrastructures
& monitoring
nd
Scenarios and use cases

In-depth analysis of Virtual


Infrastructures vs. Monitoring
By Dennis Zimmer
CEO opvizor GmbH,
VMware vExpert, VMware VCAP

Table of Contents
1. Introduction ......................................................................................................................................... 3
1.1 Virtual infrastructures are becoming increasingly complex ................................ 3
1.2 A Wide range of virtualization solutions and infrastructure components .... 3
1.3 Keeping systems reliable through monitoring ............................................................ 3
2. Operating Monitoring solutions ................................................................................................. 4
2.1 Setting the right threshold .................................................................................................... 4
3. Depth ...................................................................................................................................................... 4
3.1 Removing ambiguity ................................................................................................................ 4
3.2 The difference between in-depth analysis and monitoring .................................. 4
3.3 How to respond when problems arise ............................................................................ 6
4. A Question Of Correct Analysis .............................................................................................. 6
DISCLAIMER .......................................................................................................................................... 8

A deep analysis of virtual infrastructures & monitoring

1. Introduction
1.1 Virtual infrastructures are becoming increasingly complex
Virtualization is an indispensable part of a modern data center. Frequently, the
degree of virtualization is 90 percent or more. What formerly operated on a number
of servers today runs on a few hosts. With the high rate of virtualization and the
resulting increase in complexity, problems are more difficult to locate. It is therefore
necessary to consider how the infrastructure can be monitored accurately and how
potential error situations can be found to avoid costly errors. Unfortunately, under
certain circumstances, even minor problems can significantly negatively impact the
entire infrastructure.

1.2 A Wide range of virtualization solutions and infrastructure


components
Virtualization solutions are many: the selection ranges from suppliers such as KVM
and Citrix to Microsoft Hyper-V and the market-leading provider VMware, with its
vSphere solution. The variety of combinations with other components of the
infrastructure is limitless. Reduced to its basic functionality, each of these solutions
operates almost the same way. They mainly enable resource partitioning for optimal
and cost-effective use of physical hardware. In addition, completely new methods of
high-availability designs are possible.

1.3 Keeping systems reliable through monitoring


What about the reliability of the virtualized machines (VMs)? Are the smooth
operation of VMs and the applications running on your systems guaranteed?
Keeping track of this complex infrastructure can be guaranteed only by employing
various tools, with at least one monitoring solution serving as the base. The aim is to
be promptly notified if system loads are exceeded or failures occur. In many
organizations, failure prevention tools offer 99.9% even 99.99% reliability. Such
statistics are not possible without appropriate software and automation.

A deep analysis of virtual infrastructures & monitoring

2. Operating Monitoring solutions


Monitoring tools are widespread, such as Nagios or Icinga, Microsoft SCOM or
proprietary and application-specific monitoring tools (e.g. integrated in VMware
vCenter). They offer real-time insight on whether certain thresholds are exceeded
or if a failure has occurred. If this is the case, then the software alerts the
administrator through email or SMS sounds an alarm.

2.1 Setting the right threshold


The biggest challenge is the correct setting of the threshold value, since this
threshold determines whether an action should be performed or not. For example,
sensitive thresholds lead to many alerts and alarms, and administrators are flooded
with harmless or false messages. This causes really important messages to
sometimes be overlooked in the crowd. But what is the correct threshold for an
administrator? This must be decided based on the unique infrastructure. But of
course, recommendations and best practices exist that can be implemented and
provide guidance.

3. Depth ANALYSIS?
3.1 Removing ambiguity
An analysis is by definition a systematic study which consists of two processes, data
collection and evaluation. In particular, we consider this relationship and its effects
and interactions between the elements. In the analysis it is always about the
evaluation of the data obtained.

3.2 The difference between in-depth analysis and monitoring

Fig. 1

A deep analysis of virtual infrastructures & monitoring

On Fig. 1. you can recognize how an issue could escalate if it is not detected by indepth analysis. The time to act could be increased tremendously if a tool for in-depth
Detection has been set up in the infrastructure.
An in-depth analysis of the infrastructure is usually tested in accordance with rules,
security and best practices. It's less about the actual state of the load, but rather the
HOW, i.e. how something is configured. For example, a message such as "100% CPU
utilization" appearing without more information would not be very helpful. Here
you can already see a clear distinction between pure monitoring and analysis. You
want to know why the reported problem occurred and how it can be fixed.
Therefore, an automatic recognition after troubleshooting and recording would be
ideal.
A typical example which comes into play at each virtualization manufacturer involves
the topics vCPU (virtual CPU) and vMemory (memory which is assigned to a virtual
machine). Surely every administrator has received a request to create a virtual
machine with x number of vCPUs and y GB of RAM. But how will the administrator
take notice if the resources fulfill the requirements of the virtual machine or if sizing
is totally overprovisioned? This is where a deep analysis comes into play. It can be
analyzed using various values, where the corresponding information for resource
optimization is then displayed. For a too high number of unnecessary vCPUs can be
a performance problem on the respective host system, too. Additionally, we must
always bear in mind that a virtual machine is rarely alone it has as many systems
that can be deployed on the physical host without interfering with each other. Thus
although it may not be directly relevant, an optimally configured resource impacts
the overall infrastructure.
The added value of an analysis in automation is to screen information on system
configuration and measure the results against predefined rules. The administrator
can, of course, check such items manually against best practice recommendations.
However, this can be daunting due to the size and complexity of some
infrastructures and is also quite error-prone. According to best practices, more
components are evaluated and recommendations are made depending on the
current version. In the virtual environment, attention should be paid to how storage
and network components work together. Another popular theme is whether
clusters are uniformly configured. Through a deep analysis, the administrator wants
to be preventively informed. This also enables the ability to respond before an error
to avoid breakdowns and lags in productivity. Once you imagine that up to 512
virtual machines are supported per physical host (of course a very symbolic number),
the need to operate optimally becomes clear.
In the meantime, applications that are always running, such as SAP, Microsoft
Exchange, SQL, SharePoint, Tomcat, etc., are critical to the business. But often the
request is only for a virtual machine, without the knowledge of what might be
running. In this situation, how can a virtual machine be configured optimally for the
request? Usually not with the default values, which are at times just clicks through a

A deep analysis of virtual infrastructures & monitoring

wizard. Often it's the little things that matter, like the right selection of a virtual
network card or the correct SCSI controllers in the virtual machine.

3.3 How to respond when problems arise


The administrator receives information that an event has occurred through the
monitoring system, which must then be routed to troubleshooting. Given the large
number of complex components that are used in a virtual infrastructure,
troubleshooting is often quite difficult. Is it just storage latency problems or even
misconfigured MTU sizes on the switches? There are several tools to support the
administrator In the VMware environment. Esxtop is one popular tool. However,
using it effectively requires some know-how, especially when interpreting thresholds.
The site administrator therefore usually relies on their own initiative. Whats more,
an immediate or timely solution is needed.
The in-depth analysis is different in relation to monitoring in the way that problems
that are encountered are treated. A CPU utilization problem of a virtual machine at
100% is displayed and reported, but the administrator doesnt have the information
on why this CPU problem has occurred. In many cases, CPU limits are set in the VM
configuration temporarily, and then removing the limit is forgotten. Thus, in-depth
analysis combines a monitoring system with an appropriate expert system.
A new approach for in-depth analysis includes tools such as opvizor.
As Andreas Peetz, vExpert and blog author (http://www.v-front.de) said:
"Opvizor lets you run health checks and predictive analyses in a fully automated way.
These are derived from up-to-date rules that are centrally provided by notable
virtualization experts. Based on these "cloud rules" you can e.g. create weekly
reports that are available anytime, anywhere. This way the virtualization admin is
enabled to act preventively, but without burdening himself with maintaining
complex software, because that is implemented as a real cloud service. Only one
small local agent is needed in your environment. In a nutshell opvizor helps to avoid
many issues and outages and makes the administrator's job easier and much more
efficient. This software is definitely worth an investment!"
ALL A QUESTION OF THE CORRECT ANALYSIS

4. A Question Of Correct Analysis


It is not always easy to find THE solution for excellence for a given infrastructure.
However, you have to consider how individual software products work together best
in the area of in-depth analysis and monitoring and also what gives the administrator
a secure feeling (see also fig. 2).
Thanks to Big Data, sufficient meta-data is usually available from the virtual
infrastructure. However, these need to be properly evaluated and thats where indepth analysis enters.

A deep analysis of virtual infrastructures & monitoring

A deep analysis is the guarantee of a high-performance, secure, and error-free


infrastructure. It reduces errors and warnings in the monitoring tools and relieves
the administrator of the troubleshooting burden, allowing time for higher-value
projects.
Type
Monitoring
In-depth analysis

Use Case

Effort to Configure

uptime surveillance

high

In-depth Compliance check

low - medium

Fig. 2

A deep analysis of virtual infrastructures & monitoring

DISCLAIMER
Copyright 2014 opvizor GmbH, all rights reserved
The content and the information in this document are protected by copyright. This
emphasis, processing, distribution or duplication (copying by any means) of this work
or portions thereof, are not permitted without the consent of the publisher.
The information in this document is provided together with the VMware opvizor
analysis software.
This document is for informational purposes only. opvizor GmbH assumes no
liability for the accuracy or completeness of the information.
To the extent permitted by applicable law, opvizor GmbH provides this document as
is without warranty of any kind, including in particular the implied warranties of
merchantability, fitness for a particular purpose and non-infringement. In no event
shall opvizor GmbH shall be liable for any loss or direct or indirect damages arising
from the use of this document, including, without limitation, lost profits, business
interruption, loss of goodwill or lost data, even if opvizor GmbH has been advised of
the possibility of such damages.
The opvizor GmbH reserves the right to make changes and improvements to the
product in the course of product development.
opvizor GmbH
Schnbrunnerstrasse 218-220 , staircase A 4.04 A-1120 Vienna, Austria
UID: ATU67195304
www.opvizor.com
CEO : Dennis Zimmer

Date: May 3, 2014

A deep analysis of virtual infrastructures & monitoring

Вам также может понравиться