The Sightline™ Power Agent™ for Unisys ClearPath OS 2200 Systems resides on the ClearPath host and is easily configured to collect data on various key system components and user-specified programs and groups of programs as single measurable workloads. The data is collected on a regular basis, stored on the host, and sent via TCP/IP to Sightline Enterprise Data Manager (EDM) for alerting, analyzing, displaying and reporting.
EDM enables IT teams to continuously monitor the entire system in real time in order to proactively predict performance issues and prevent unplanned downtime and data loss. Its highly interactive, visual tools are used to achieve results in minutes, accelerating discovery and investigation within any IT infrastructure.
In addition to the Sightline Power Agent for ClearPath OS 2200 Systems, Sightline offers Torch Capacity Manager. Torch is a performance management system used for long-term workload trend analysis and capacity planning for ClearPath OS 2200 systems. Torch combines key performance data from multiple sources into one comprehensive integrated performance database. It can produce graphs and statistical reports that summarize and display performance data according to specific needs, while also providing the information necessary to make sound decisions concerning system performance and capacity requirements.
The following list describes the metric groups collected by the Sightline Power Agent for ClearPath OS 2200 Systems.
The System Instrumentation Package (SIP) is a standard part of the OS-2200 operating system. SIP monitors and collects performance statistics about the operating system, workloads and major pieces of hardware on the system. A comprehensive set of statistics is supplied. Sightline provides statistics on IP, Memory and I/O utilization, XPC, Gate Activity, IPM, Expool, Dynamic Allocator, Dispatcher, ER activity, and hundreds of other critical performance metrics. The SIP data source is included with the Sightline OS 2200 Power Agent.
The System Log provides detailed TIP, demand, and batch resource usage statistics by individual run, program, and terminal. Sightline takes advantage of data available through INSP$ calls so you get the information you need to link performance problems back to the user and transaction. Additionally, disk errors and security violations are monitored so that failing hardware or unauthorized usage can be detected before major system outages occur.
TIP KONS is an area of memory set aside by the Transaction Interface Processor (TIP) to record the performance and behavior of a TIP transaction environment. The global metrics from System KONS can be used to determine the need to look at individual transaction statistics. There are eight areas of performance information available within the KONS Interface Agent: Transaction Response Times, Transaction Rates, Transactions Active, COMPOOL, Transaction Errors, FCSS, Transaction Scheduling, and Timer Statistics.
MAPPER data collected by the Power Agent is divided into six functional groups: Communication, Mass Storage, Memory Pool, Report Access, User Activity, and Workload. Workload data provides Utilization, Activity Counts, and Response Time Statistics by MAPPER function, department, user, run and type.
The Master File Directory (MFD) can be queried to obtain information about the mass storage of a system. Sightline provides statistics about the availability (MFD K-Trks Avail) and the reliability (MFD KNonErr Refs) of each disk drive and logical channel. Statistics are also provided summarizing these statistics for all cache, non-cache, fixed, removable, and for all disks. Additionally, Sightline can be configured to produce MFD statistics by qualifier, project-id and account number.
The Universal Data System (UDS) measures transaction volume, queuing and errors in any of the databases managed under the UDS umbrella. UDS data is gathered from the published UMON interface or from the ACT, QAD and BST Tables within UDS. Hit rates can be determined, queuing problems can be quickly identified and overcome and space utilization can be optimized. Sightline collects data for three functional areas: Transaction Activity and DMR Slot Usage, Buffer Usage plus Rollbacks, and Queuing by queue reason.
Transaction Performance Monitor (TPM) audit trails and HVSTATs are used in place of TIP Logging at some sites. The statistics reported are similar to those described above for TIP Log, with the addition of several statistics unique to HVTIP. With Sightline, the TPM Log records can be grouped into workloads characterized by Program Name. HVSTAT records can be grouped by Program Name or Initial Control Program, User Word030, or LINC ISPEC.
Step Control metrics record tree usage and queuing by node and priority. Message queuing can be isolated in the input tree – queuing that will not show up directly as increased response time. Step Control identifies transaction being queued before they’re scheduled. (System Log, TPM/HVSTATS, and UDS data identifies transactions being queued after they have been scheduled.) Sightline collects statistics for all configured priorities and nodes and displays input messages active and queued, total and maximum allowed; and output messages active, queued and total.
CMS or SILAS data provides an event-by- event description of everything that went through the communications system. Data collection from multiple instances of CMS is possible.
The Display Processing System (DPS) allows application programmers to define and use screens for their programs, to separate screens from other parts of the application. Sightline provides statistics on the DPS system including message activity, errors, I/O rates and utilization, and response time.