WHITE PAPER Unified Monitoring Drives High- Performance Business Results
Table of Contents EXEC SUMMARY... 1 INTRODUCTION... 1 THINK BEFORE YOU BUY... 2 The Pitfalls of Silos...2 Monitoring Tools: Less is More...3 Legacy Tools Not up to Today s Workloads...3 What CIOs want in 2015...3 SERVICE-CENTRIC UNIFIED MONITORING... 4 Rallying Diverse IT Stakeholders around Common Goals...5 Tool Consolidation...6 Support for Modern Technologies and Architectures...6 Integration with Domain Specialized Tools...7 SUMMARY... 7
Exec Summary Before acquiring more domain specific monitoring tools, such as APM (Application Performance Monitoring), IT departments should take the time to review their monitoring strategy to ensure a unified approach to monitoring. Although it is generally agreed that there is no single monitoring tool that can satisfy the diverse needs of all stakeholders, there are strategies that can link best-of-breed tools in both domain specific monitoring and service-centric monitoring to promote cross-team collaboration, align monitoring and management activities with the business and reduce the number of monitoring tools in use. Fewer tools translate into faster problem identification, more efficient resolution, and more comprehensive infrastructure management. Introduction Service-centric unified monitoring can: Reduce business-impacting downtime associated with performance and availability issues Unify disparate monitoring teams and provide a business aligned priorities for collaboration Simplify and reduce the operational effort of monitoring Reduce the number and excessive cost of many redundant monitoring tools Replace outdated tools with lower cost monitoring solutions for modern architectures How do IT budget owners determine the best investment strategy when it comes to monitoring? Today, most have an abundance of monitoring tools covering Infrastructure Performance Monitoring (IPM), Network Performance Monitoring (NPM) and to a lesser extent, Application Performance Monitoring (APM). Many IPM tools have outlived their usefulness and can t handle virtualization and modern cloud architectures. Their agent-based data collection functionality makes deployment and operation unnecessarily complex and older agents can become potential security risks. New tools seem to be the only solution. New monitoring tools must transcend individual domain teams. Without an overall performance view, there will be more frequent and longer delays in resolving performance issues that impact the business. Consider the parable of the elephant and the blind men: One day three blind men encounter an elephant. To discover what type of animal it is each feels a part of the animal, but only one part. One feels a leg and says the elephant is like a pillar; another feels the tail and says the elephant is like a rope; the third feels the trunk says the elephant is like a tree branch. They disagree as to what animal it is and they begin to argue, shouting, That s not an elephant! Yes it is! No, it is not! and so on. When a sighted man walks by and sees the entire elephant, he helps them begin to work together to see the full elephant, using his gift of sight to create the whole. The elephant is a performance problem to be solved. Each of the blind men represents a different set of monitoring stakeholders who can t see beyond their own silos. The result? Finger pointing and excessive downtime. An IT organization s trusted-advisor role should include both a big picture view of performance and specialized capabilities for individual stakeholders. 1
Think Before You Buy The first step is a closer look at the current monitoring tool strategy and portfolio to understand the shortcomings of silo monitoring, as well as excessive and outdated tools. Are new point tools truly filling a gap in monitoring coverage or merely adding additional capabilities that already exist elsewhere? For example, can you justify the cost of APM tools for every application you have? Is it even possible to use APM on every application you own? Closer examination may allow you to adjust your ratio of spend between point tools and tools that can satisfy the monitoring needs of multiple teams, increase cross-team collaboration and reduce overall operational costs. The Pitfalls of Silos Typically three functional areas comprise the majority of required performance and availability monitoring (see Figure 1). Application support stakeholders manage the application and database functions. IT operations stakeholders manage converged infrastructure, physical and virtual compute, cloud, and storage functions. Network stakeholders manage network performance. Each team uses different tools. The moment a problem arises, disconnected tools and views lead to finger pointing between the groups and excessive delays in problem resolution. Without cross-team collaboration and prioritization of effort, monitoring teams measure success in IT terms such as server uptime, response time, and database transactions per minute, rather than business metrics. FIGURE 1 TRADITIONAL SILO DOMAIN MONITORING 2
Performance and availability monitoring strategies require a unifying solution that enables disparate stakeholders to collaborate in a natively integrated fashion. Budget owners should resolve the siloed monitoring problem by acquiring tools that can link business-focused IT services to the underlying IT infrastructure. A service-centric monitoring tool aligns the monitoring stakeholders to business priorities by consolidating redundant capabilities from different tools into a single platform. These tools increase collaboration between the teams through shared data, and integrated, role-based views. Monitoring Tools: Less is More Most monitoring portfolios include a proliferation of tools due to the failure of system management frameworks. Failing to meet the needs of diverse monitoring stakeholders leads to independent tool purchases across multiple groups. As tools continue to proliferate within IT organizations, overall monitoring tool Total Cost of Ownership (TCO) steadily increases. In addition, redundancy of monitoring capabilities across tools also increases total costs (see red lettering in Figure 1). Unable to correlate data between redundant tool capabilities, questions of accuracy and increased mean-timeto-resolution of performance problems are common. While it s not possible to consolidate all monitoring tools into one solution, it is possible to significantly reduce the number and cost of monitoring tools. Solutions that leverage agentless collection technology and intelligent, role-based displays satisfy a much wider range of buyers for significantly less. Legacy Tools Not up to Today s Workloads New technologies and architectures make legacy monitoring solutions into barriers instead of enablers. Monitoring solutions must work with distributed workloads scaled out across hundreds or thousands of public and private servers. Virtualization, cloud and agile software development are outpacing older monitoring technology. Modern monitoring solutions continuously discover the production environment including components, their relationship to each other, and the business services they enable. What CIOs want in 2015 Reevaluating your monitoring strategy and tool portfolio could not be more relevant. In a recent survey conducted by Gartner, CIOs rated modernization of infrastructure and data center and BI/analytics as their top-two spending priorities. Modernization of the data center will bring new capabilities and benefits at the cost of increased complexity in infrastructure and application architectures. Monitoring strategies and tools need to be up to the task. Advances in IT operational analytics provided in new monitoring solutions can mine a wealth of performance data that have specific business use. A recent survey from the National Association of State Chief Information Officers (NASCIO) corroborates the Gartner findings regarding data center modernization and adds a priority. In this survey, two of the top five priorities for CIOs in 2015 underscore the need for practical and affordable approaches to performance and availability monitoring. Consolidation/Optimization: centralizing, consolidating services, operations, resources, infrastructure, data centers, communications and marketing enterprise thinking, identifying and dealing with barriers. Budget and Cost Control: managing budget reduction; strategies for savings; reducing or avoiding costs; dealing with inadequate funding and budget constraints. 3
Most enterprises must engage in a serious re-evaluation of their monitoring strategies. Tools need to unify monitoring activities across disparate IT stakeholders to reduce or eliminate the economic impact of availability and performance issues. Consolidation of monitoring tools into simple-to-deploy-and-use solutions to increase collaboration, lower TCO and improve the quality and performance of delivered services can accelerate 2015 CIO priorities. Service-centric Unified Monitoring Service-centric unified monitoring (UM) is a practical and affordable monitoring strategy that: Provides a holistic view of performance and availability monitoring to align different stakeholder perspectives to business priorities Consolidates monitoring capabilities into fewer, simpler, cheaper tools Integrates with specialized monitoring tools needed by individual stakeholders A UM strategy enables maximum collaboration during incident and problem management across IT operations, application support and network management teams. A UM platform converges IT stakeholders onto services critical to the business for conducting commerce, reducing risk or supporting internal initiatives. Though not a one-size-fits-all solution today, service-centric unified monitoring tools can resolve two critical monitoring issues that plague IT: Insufficient IT stakeholder collaboration on business impacting issues Too many outdated monitoring tools that are expensive to own and manage In contrast to Figure 1, Figure 2 illustrates the same three functional areas of availability and performance monitoring (i.e. Application, Infrastructure and Network). In this example, service-centric unified monitoring principles have been applied. In area 3 of the diagram, a service model with service impact and analysis is now central and visible to all the stakeholders managing these functional areas. Application support, IT operations and network teams are aligned and collaborate on a common set of business-aligned services. Event management has been combined into a single event console. Log search and analytics have been added to supplement the infrastructure monitoring domain s switch to agentless collectors. In areas 1 and 2 of the diagram UM has consolidated a number of capabilities between domains into a common shared tool. Specialized capabilities remain within the domain- specific tools for use by individual domain teams. These same tools integrate into the common event management and service-impact models to correlate domain-specific conditions back to the business services they impact. Additional monitoring capabilities not present in older solutions (indicated in red) are available to monitor modern architectures and technologies. 4
FIGURE 2 SERVICE-CENTRIC UNIFIED MONITORING Rallying Diverse IT Stakeholders around Common Goals The heart of today s service-centric unified monitoring solutions is a service impact model. See Figure 2, diagram area 3. This model allows IT stakeholders to work together to define critical and non-critical business services that must be maintained. Business critical service containers are typically mapped to key business transactions and their supporting application components by development or DevOps teams. IT operations and network teams define dependency rules to map the supporting infrastructure and network elements into the same service container. The result is a key service that rallies diverse IT stakeholders around a common goal. Each stakeholder uses their expertise, knowledge and domain-specific tools to maintain service performance and availability. Service containers may also be used by individual stakeholders to help prioritize monitoring and management goals within their team. Once defined, UM 5
solutions continuously discover configuration additions and changes to the physical and virtual environment to ensure accuracy of service container dependencies. The collaboration benefits of the service-centric unified monitoring strategy include: Faster problem impact assessment and problem resolution More efficient collaboration prioritized to business goals Improved visibility and increases service delivery quality Tool Consolidation UM solutions have the unique potential to eliminate multiple tools from an environment by consolidating functionality into a single platform. Significant savings can be realized up front as the solution is deployed, or phased in over time. Redundant capabilities can be replaced with lower-cost agentless collection technologies. Eliminating legacy tools with a high TCO or deferring additional point solution purchases to fill gaps in coverage can reduce costs. In Figure 2, areas 1 and 2 a number of monitoring capabilities can be consolidated into a UM solution. The tool consolidation benefits of the UM strategy include: Significantly reducing the number of tools required Lowering monitoring costs by replacing expensive legacy tools Support for Modern Technologies and Architectures Unlike system monitoring frameworks and older domain-specific tools, UM tools focus on simplicity at all levels. Agentless monitoring has replaced older agent-based capabilities. These new capabilities exploit a new generation of APIs to eliminate the management overhead, complexity, rigidity and security vulnerabilities of agent technology. Agentless technology is much easier to use in the cloud and can significantly reduce operational costs and. For example, hundreds of virtual machines can be instantiated in minutes. Agentless technology uses existing APIs to collect monitoring data to avoid the significant operational and licensing costs associated with deploying agents. UM solutions operate at higher levels than domain specific monitoring, allowing them to provide discovery and topology analysis to support service impact and analysis. Log file analytics correlate events and aid troubleshooting across hundreds of devices in complex environments and replace the need for deep-dive agents in all but the most specific cases. One of those cases, application transaction tracing, allows developers to follow code-level transactions through their custom applications. Tracing requires resource-intensive agents. APM tools that have this capability are complex to use and extremely expensive to purchase. Transaction tracing is limited to custom applications built with languages (e.g. Java,.NET, etc.) that agents can understand. Although essential for some applications, APM is too costly to use on any but the most critical business applications. According to Forrester Research companies only spend 24.6% of their software budget on custom applications. The remaining applications need performance monitoring coverage that can be provided by UM solutions. Advances in service-centric unified monitoring coupled with log analysis can replace the need to use APM tools on all but the most critical apps. Given the relatively low cost of UM solutions, this can lower licensing and operational costs significantly. 6
Integration with Domain Specialized Tools UM solutions keep IT stakeholders focused on what matters to the business, and allow different stakeholders to use the capabilities best suited for their role. The centralized monitoring and alerting afforded by UM s service impact and analysis capabilities enables stakeholders to simultaneously monitor overall service delivery performance while performing their domain-specific duties. For example, application support teams in a DevOps role can use the real user experience monitoring and transaction tracing capabilities of APM to troubleshoot performance issues related to incorrect or inefficient code. This allows them to remedy as quickly as possible, unencumbered by traditional IT processes. Alerted to a business impact, the network team may opt to focus deep-dive troubleshooting efforts using their domain specific NPM tool. Centralized IT organizations and NOC teams looking to become more business-savvy can adopt UM capabilities like synthetic user experience monitoring as a simple way to begin the transition away from purely infrastructure-centric operations. Finally, UM solutions easily integrate with other third-party tools such as ticketing systems (i.e. ServiceNow) or other specialized monitoring tools through off-the-shelf or custom integration modules. Summary Before acquiring more domain-specific monitoring tools, such as APM, IT departments should review their monitoring strategy to ensure a unified approach. While today there is no single monitoring tool that can satisfy the diverse needs of all stakeholders, there are strategies that can link best-of-breed tools in both domain-specific monitoring and servicecentric monitoring to promote cross-team collaboration, align monitoring and management activities with the business and reduce the number of monitoring tools. Many organizations have successfully funded the costs of a service-centric unified monitoring solution simply by decommissioning existing tools and eliminating their associated costs (see Making the Business Case for Unified IT Monitoring Solutions from Zenoss). i Gartner Group, Tech Go-to-Market: 2014 CIO Survey Indicates Implications for Software Providers, June 3, 2014 ii NASCIO.org, State CIO Priorities for 2015, October 30, 2014 iii Forrester Research, Ten Myths And Realities of the Software Market in 2013, August, 2013 11305 Four Points Drive Building 1, Suite 300 Austin, TX 78726, USA Phone: +1-512-687-6854 Toll-free: +1-888-936-6770 www.zenoss.com Zenoss and the Zenoss logo are trademarks, or service marks of Zenoss, Inc. All other trademarks listed in this document are the property of their respective owners.