How VMware monitors performance issues

The opacity of today's virtualized data centers makes traditional monitoring methods and enterprise IT departments lack a tool that allows them to determine what causes the application to slow down through the association of Applications and Network architectures.

However, the increased number of virtual machines and the complexity of workloads makes it difficult to diagnose whether application services and network performance are related to virtualized environments, where there is a lack of visibility and effective tools to determine what causes applications to slow down, through the Association of Applications and Network architectures.

The basic components of a data center network infrastructure (CPU, memory, network, and appliances) have their own management tools, but the existing tools do not see the performance of the entire virtual data center, or provide the relevance of the information behind a particular application's performance, Thus, there is a lack of effective systems to solve problems from resources or applications faster and more efficiently, and it takes a lot of time to solve the problem. In particular, the key service host services, such as DNS host, RADIUS authentication host, Mail gateway and other important equipment is reaching. Once there is a problem with the virtual center, managers are at a loss to find the underlying problem quickly and efficiently.

Addressing the root causes of today's applications and virtualized infrastructures requires visibility into application services, networks, and virtualized architectures that apply to the various departments of the Information Center. Service outages and slow-processing issues can cause huge losses to the company, and addressing or preventing such problems is the primary task of the Information Center. The information team needs to work together with interrelated data collaboration to avoid the blame and try to find the root cause before the failure affects the network service.

650) this.width=650; "src=" http://netmanager.com.cn/upload/img/201702081538062743.jpg "style=" width:701px;height : 519px; "/>

Second, the solution

Fully deploy the Uila VMware Virtualization Performance monitoring system with each VMware Data Center, with the following features:

1. Automatically explore the relevance of application service performance to resources and provide a view of the underlying causes of performance degradation. Contains CPU, memory and storage devices.

650) this.width=650; "src=" Http://netmanager.com.cn/upload/img/201702081539187595.jpg "width=" 626 "height=" 298 "/ >

2. Provide a unique virtual Smart tap for deployment to analyze network traffic between hosts and online response times. No physical wiring or installation of agent-side programs. Can be installed on each virtual switch or network group port of Vcenter.

3. Installation of the system does not need to stop the original virtual machine, the actual online services of the host without stopping the service can be the system.

4. The system uses the Big Data Analysis database platform to record and analyze the APPLICATION/NETWORK/STORAGE/MEMORY/CPU data statistics of the virtualized data center for a long time.

5. The system application performance through deep Packet inspection mode monitor.

6. You can collect information such as Vapp, VM topology, computer, Storage, performance metrics in vcenter.

7.Virsual Smart Tap collects and analyzes information such as application response time, network performance Metrics, network throughput, flow data, and more.

8. Quickly identify the root cause of virtualization performance and link VMware's knowledge base to provide solutions-related information.

9. The system provides baseline basic line automatic learning function, provides the warning to identify anomalies and trends, the system can define the first one hours, the day before, the previous week and a specific period of time as the Baseline Foundation.

10. With the dashboard function, the data can be presented in real-time for one hours in the virtualized datacenter, and color grading is critical (Critical), significant (Major), minor (Minor) and normal Managers can monitor the dashboard status in real time to handle every exception event.

650) this.width=650; "src=" Http://netmanager.com.cn/upload/img/201702081541274515.jpg "width=" 641 "height=" 369 "/ >

11. With application topology, provide a view of network interaction between virtual hosts. Rendering the resource and network efficiency of each virtual host deteriorated, and the Internet connection is normal and the response time too long caused the service interruption phenomenon.

650) this.width=650; "src=" http://netmanager.com.cn/upload/img/201702081544082088.jpg "style=" width:712px;height : 334px; "/>

12. Provide the network traffic distribution view, analyze each host, virtual network switch, Network group port and virtual host network traffic, and according to the port use traffic size display.

650) this.width=650; "src=" Http://netmanager.com.cn/upload/img/201702081544517972.jpg "width=" 712 "height=" 263 "/ >

13. Provide a virtual data center CPU Usage Distribution view, respectively, according to the data center, cluster, ESX host and Virtual Host CPU usage usage analysis.

650) this.width=650; "src=" http://netmanager.com.cn/upload/img/201702081545522545.jpg "style=" width:704px;height : 451px; "/>

14. Provide a virtual data center memory Usage distribution view, respectively, according to the data center, cluster, ESX host and virtual host memory usage usage analysis.

15. Provide storage usage distribution view of virtual data center, can be analyzed according to physical disk, Datastore, ESX host, virtual host and Virtual disk drive usage usage.

650) this.width=650; "src=" Http://netmanager.com.cn/upload/img/201702081546336815.jpg "width=" 701 "height=" 213 "/ >

16. Provide historical analysis data query, the data center, cluster, ESX host, Vapp and virtual Host network layer, CPU, memory and storage space data query. Response time queries for alerts and application services are also available for virtual hosts.

17. With the report output function, the application efficiency and resource utilization rate of virtual data center can be manually according to the Set time section or the periodical output Daily, weekly and annual report. According to the report, the manager can perform resource scheduling and trend analysis.

18. Provide the system log, can query the system operation and other related records.

19. With Virsual Smart tap simple installation setting interface, the manager can be based on the environment for VST installation and control management.

20. VM to VM Packet capture is available through virsual Smart tap. (need to be deployed with the Uila Viewer virtual machine).

21. Provide the capture packet and record it into the Pcap file save function.

22. Provide specific monitoring resources function of certain important virtual machine, and present the health condition of each CPU, MEMORY, DATASTORE, Network and application agreement.


III. Programme structure

650) this.width=650; "src=" http://netmanager.com.cn/upload/img/2017020815483566.jpg "style=" width:709px;height : 453px; "/>

Uila Management Analysis System (UMAs)

At the heart of Uila's virtual architecture is a large data storage and analysis engine designed to be fully scaled out (scale-out) to accommodate the deployment of large data centers with thousands of servers, to scale in (scale-in) to record data at high resolution, maintain historical data, while maintaining real-time response. Built-in redundancy provides high availability, eliminates downtime, and lowers maintenance overhead. The analytics engine is the brain that correlates application and infrastructure performance metrics to pinpoint the underlying causes of infrastructure and application performance degradation, which is patent pending. Trend reports based on historical data help identify hotspots for infrastructure resource use and help maintain optimal network application performance. Uila's Control Panel provides a simple and powerful way to view the results of the analysis engine, showing the health of networks, computing, and storage devices in applications and virtual infrastructures.

Virtual message Controller (VIC)

The virtual information controller (VIC) is a data center virtualization management System (such as the VMware vcenter) integration pipeline. Vic retrieves your infrastructure configuration as a template for establishing Uila monitoring domains and simplifying deployment. Vic collects network, storage, and compute performance metrics maintained by Vcenter, combining it with all application metadata deployed with VST, and then securely transferred to Uila's data storage and analysis engine, either in front (on-premise) or in the cloud.

Iv. Advantages of the scheme
1. Integration of analytical data into an operating interface
• Customizable application protocols and virtualized resource health dashboards that combine data to identify meaningful relationships between data, enabling performance indicators to react to performance degradation early.
• Intuitive design allows non-professional operators to easily pinpoint the root cause of slow application service response times.
• An analysis tool that combines application topologies, usage analysis, CPU utilization, memory usage, and appliance usage to uncover how virtualization resource allocation impacts application service performance.
• New and intelligent self-learning baseline technology to establish monitoring thresholds to reduce false positives.
• Integrate control platform and network Operation Center alarm information for early warning.
• A report that can be customized to meet the service level agreement; SLA.

2. Deployment mode and data monitoring analysis designed for virtual data center
• Leverage big data technology to scale to the largest data center.
• Low-energy Virtual Intelligent Packet Analyzer-no agent is installed in the operating system.
• Collect details of application response times and key infrastructure performance metrics.
• Use embedded deep packet Detection (DPI) technology to identify more than 4,000 unique applications.
• Seamless integration of virtual information Controller (VIC) with virtualized management systems such as VMware and vcenter

V. Programme benefits
1. Make the data center "visible" virtualized, complex network infrastructure, reducing the time it takes to resolve application outages from hours to minutes, and proactively adjust the performance of your business-critical applications by preventing disruptions resulting from the use of too-large network infrastructure resources.

2. Improve teamwork: In the past, the management of network management and application management and virtualization infrastructure are independent, when there are application and network problems, it is difficult to clarify whether directly or indirectly with the resources and efficiency of virtualization, through a comprehensive virtualization efficiency solution can eliminate the gap between the communication between departments, Improve the efficiency of collaboration between groups.

3. Clarify network and application effectiveness and virtualization issues: visualize virtualized infrastructure, reduce application downtime from hours to minutes, prevent outages and service disruptions that may be caused by excessive infrastructure stress, and help information center operators proactively adjust the performance of business-critical applications.

4. Prevent application performance issues: Establish performance benchmarks that can be used for long-term observation and evaluation of network applications and service demand trend planning. and can master the non-essential application service consumption, improve overall service quality and cost-effectiveness.

5. Streamline information management costs: Through the software deployment method to collect traffic and packet information, without the need for additional hardware equipment, more simple and effective deployment of the deployment can save the cost of operations and tools.

How VMware monitors performance issues