Managed Elements

Adapters, Managed Elements and Templates

Adapters leverage investments in existing monitoring solutions to collect performance metric data from existing commerical or custom monitoring data sources. Adapters are deep performance data integrations that allow the real-time collection of raw data at the lowest possible granularity. The real-time data collection ensures accurate contextual and real-time analysis for the generation of Trusted Alarms. Netuitive’s adapters are non-intrusive and have a very low impact on monitoring system architecture. Netuitive does not deploy any new agents; instead it leverages performance data already collected by existing third-party monitoring solution. The Netuitive adapters are extremely lightweight and will not consume extensive network bandwidth, nor affect the workload.

Data collected from Adapters are processed by Netuitive best-practice Templates.  These templates leverage both expert knowledge and statistical techniques to determine which metrics are the best candidates to characterize statistical behavior of servers, clusters, transactions, or any other physical or conceptual IT component. Customers view Netuitive templates as extremely valuable and a real advantage of Netuitive over solutions that have users start from scratch to determine the most relevant performance metrics. These templates represent the statistically proven and most relevant Key Performance Indicators (KPIs) or metrics for a given environment (e.g. UNIX, Oracle, Exchange, or Weblogic). Each template's pre-defined set of KPIs are then used by Netuitive’s Behavior Learning Engine™ for generating health indicators and Trusted Alarms. Customized templates can also be built using Netuitive Integration Studio.

Managed elements represent what you monitor in your IT infrastructure.  Each managed element has a corresponding template in Netuitive which essentially represents a model for the performance of an IT component such as servers, clusters, virtual machines, virtual data center, disks, transactions, JVMs, or even entire applications such as Microsoft Exchange and mission critical business applications.  Base on analyzing the performance of managed elements, Netuitive can generate simple performance indices such as Health, Workload, Alarm state and Availability.  In addition, one can drill-down into Managed Element details where attributes, correlation and performance data can be browsed and analyzed.

State of Managed Element: Health, Workload, Capacity, Alarm and Availability Indices

The state of each managed element is determined through five automatically calculated indices: health, capacity, workload, alarm, and availability.

The health score of a managed element is a composite index with a value between 0 and 100 that reflects the overall status of a managed element from the following perspectives: performance, availability, capacity and workload. The health accounts for the workload by using the probability that the workload will exceed 95% in the next monitoring cycle (typically 5-15 minutes), the alarm index that represents how a managed element is behaving as expected, the capacity index indicates how much resources are being consumed and the availability whether the ME is up or down.

The workload index of a managed element also has a value between 0 and 100 and is a composite measure of the amount of resources expended or work performed by a particular managed element. Each managed element has in fact multiple workload indices—one for each resource defined in the matched template. The formulas used to compute the workload indices are defined in each template and vary based on the technology and the integration type. They are computed and based on specific individual metrics that are defined in the specific template.

Unlike capacity, the workload accounts for resource sharing overhead; as such, disk read, disk write, disk queue, disk i/o, context switches, paging, etc. These key performance indicators are being used in the workload calculation in order to offer a more meaningful measure of resource consumption and capacity. The capacity index may indicate that resources are left; however, it does not mean that it can be consumed. For example, processing cycles may be available but cannot be used because too much paging and time is being spent for memory access, or the system is thrashing because of context switch and no work can be performed. The alarm index is a representation of how faithful the managed element current behavior is to its historical and expected behavior learned from the data by the Behavior Learning Engine. It also accounts for manually define policy thresholds and service level objectives (manually or automatically defined) based on static and statistical thresholds. Alarm range can take four values: warning, minor, major and critical. The alarm index is a representation of the average severity of an incident composed of consecutive alarms. The availability index is the rudimentary index that represents a system to be up or down. Netuitive also has an availability warning status that indicates that a system is at risk when too many of its components are down (e.g., fail over cluster running on its backup server).

Integration Interfaces & APIs

Netuitive Integration Platform Interfaces (APIs)

Netuitive predictive analytics plays a key role in the IT operations eco-system and, as such, the software has a number of programmatic interfaces and tools that help it fulfill its role.  Netuitive's Open Platform initiative is our commitment to ensuring greater customer value by creating an extensible platform that allows users to access data from anywhere to customize their application of Netuitive analytics and performance models.  The added benefit is that customers are not locked in to any large framework solution fromm any single vendor. Integration Studio is a set of development tools that support this initiative.  More specific use cases for Netuitive's Open Platform initiative are detailed below.

Monitoring Data Collection

• Standard adapters to collect Key Performance Indicators (KPIs), events, and system attributes from 3rd party monitoring tools supported by Netuitive.
• Integration Studio is a set of tools that enables users to create custom templates and adapters for integrating with Netuitive.

Service Definitions

• Configuration Management Database (CMDB) integration is via a Web API using CSV or XML to define service configurations.  Direct integrations are also available to certain commercial CMDBs (e.g. HP uCMDB).

Administration and configuration

• User and security management can be done programmatically via the Web API in addition to Netuitive’s graphical user interface.
• The Web API also supports programmatic definition of managed element groups, and policies.

Incident Management

• For incident management, the Web API supports enrichment of the Netuitive alarm knowledge base as well as alarm export to centralized 3rd party alarm consoles. The enriched alarm information can also then be forwarded to Service Desk applications to add diagnostic information to service tickets.

Resource Management

• Netuitive can integrate to Service Director or Run Book Automation applications using triggers or analytics results to manage resource management / allocation decisions.

Reporting

• In addition to the standard reports and specialized report packs available in Netuitive, there are a number of ways to generate custom reports.
• JRXML (Jasper Report XML) is an open source reporting library that can be used to generate custom PDF and CSV reports.
• Detailed performance and other data can also be extracted from the Netuitive Performance Management Database (PMDB) and exported to custom databases, data warehouses, or 3rd party Business Intelligence tools like SAS, Business Objects, Cognos, etc.

Performance Management Database (PMDB)

Performance Management Database (PMDB)

To perform its cross-silo analysis, Netuitive collects performance metric data and metadata from many different data sources. From the raw data it creates calculated metrics (e.g., workload, health, capacity indices), dynamic thresholds (historical baseline, contextual baseline, forecasted baseline), roll-up statistics (daily, hourly, 5/95th percentile, average, standard deviation, alarm counts, deviation counts, etc.) trends (1, 2, 3 and 4 weeks, 1, 2, 3 and 6 months into the future).This makes the resulting database the most comprehensive repository of performance data in your enterprise – the PMDB.

The PMDB is a repository for Netuitive’s analytical results for use by its User Interface, reports, dashboard and search engine but can also be a source of performance data for strategic analysis using tools like SAS, Business Objects, etc. – eliminating the need for multiple integrations from these tools to raw performance data from existing monitoring tools and agents.

When configured solely for real-time analytics, the PMDB retains only the analytics calculations necessary and does not store raw metric data for any extended period of time. This minimizes data base storage requirements.  The PMDB is usually configured to retain data for a longer time period to support historical reporting and long-term analysis as well.