An overview of the metrics collected by the Sightline for Unisys MCP PPA software is provided in the following paragraphs. In addition to the PPA metric groups, optional Interface Agents can be included so that you have the information you need about all of the critical applications running on your system.
In addition to the Performance Power Agent, Sightline offers the Capacity Power Agent and the Sightline Workload Analyzer for ClearPath MCP systems.
The Sightline Capacity Power Agent (CPA) is an advanced Sightline data consolidator and capacity database manager based in the Unisys MCP environment. It provides an automatically managed repository of historical performance data for the Unisys host. The CPA database resides on the Unisys MCP host and is populated with interval data (30 or 60 second samples of performance data) from the MCP and all Interface Agents as it is collected. Transparent to the user, the collected data is automatically rationalized to 15 minute, hourly, daily, weekly, and monthly summaries as well as into user definable custom shifts for long-term storage, retrieval and analysis. Management of this performance and capacity database is automatic, with data aged out of CPA based on retention specified by the user.
If both Sightline PPA and CPA are resident on the same platform, then both the MCP performance data collection and the Interface Agent data collection are integrated and shared between the two Power Agents to improve efficiency and reduce collection overhead.
Sightline CPA is seamlessly integrated with EDM for alerting, analysis and reporting but it also comes with a fully functional host-based reporting mechanism. A parameter driven utility on the host allows the user to format and schedule printing of host based tabular reports and to export data to other analysis tools such as Excel™.
The Sightline Workload Analyzer (SWA) is a stand-alone facility that integrates with the other Sightline products by providing detailed task level performance statistics for all task executions. SWA offers the capability to monitor critical task/job execution, track batch cycle execution, perform deadline-processing analysis and conduct exception analysis. SWA automatically maintains a database containing up to several years of records of individual task and program executions integrated with snapshots of resource usage data.
This comprehensive and efficient data storage facility permits easy reporting and analysis of system resource usage by program for both active and completed programs in a period of interest and facilitates daily, weekly, monthly, and longer term reporting of batch cycle and deadline performance. The SWA Monitor automatically collects and stores data without the need for operator intervention and may be configured to automatically backup data and remove data after a specified duration. The SWA Reporter interfaces to the PPA/CPA Parameter file to extract Workload and Billing Account definitions to facilitate ease of reporting.
The following list describes the metric groups collected by the Sightline Performance Power Agent for ClearPath MCP Systems and PPA’s Interface Agents.
Sightline reports the amount of available and in-use memory, plus statistics that profile the Overlay activity of in-use memory and the users of the memory resource. This data helps you detect memory constraints causing performance bottlenecks.
Sightline reports I/O data by I/O Processor (IOP), Subsystem, Unit, and Family. Performance measures include Percent Busy, I/Os per Second, Average Block Size, Seconds per I/O, and Average Queue depth. This allows for elimination of I/O bottlenecks and optimal allocation of I/O activity to improve system throughput.
For each Network Processor (NP or ICP), Sightline reports I/O traffic rates for reads and writes, average I/O times, average block (message) size, queue depth, percent busy, and other key metrics. This data, when supplemented with TCP/IP data, can help you improve the efficiency of communications line handling and the system interface to your LAN/WAN.
Sightline allows workload analysis at two levels: individual online transaction types and program workload classes. Each online transaction type is measured in terms of Rate per Second, CPU Time, I/O Time, Response Time, Queue Time, etc. For each user-defined program workload class, Sightline reports the Mix Count, CPU, and I/O utilization, memory utilization and ReadyQ Percent. Different performance objectives can be set for each workload class.
Sightline EDM’s Event data display reports on a variety of “worst offenders” including busiest TCP/IP LANs, busiest BNA connections, busiest database structures, busiest online transaction processors and the programs or tasks that are consuming the highest processor, I/O, and memory utilization, all sorted by user-selected resource criteria. Data can be viewed in real-time or after the fact. The Event data feature allows you to diagnose system communication bottlenecks and resource consumption problem applications at the program level, data structure or connection level.
Application programs can easily insert data into the Sightline data stream. The application metrics will then be stored in the Sightline host Trace Files and the Sightline Capacity Manager, and subsequently forwarded to the Sightline EA/V workstation for analysis along with the Sightline performance metrics. This permits meaningful application metrics (for example, order counts and values from an order processing application) to be coordinated with the Sightline-collected application performance and resource usage metrics.
The Sightline Interface Agents for Unisys MCP systems collect performance metrics from a variety of non-MCP components of the Unisys MCP environment. Interface Agents provide data in addition to the PPA’s system data and also to Sightline Capacity Manager, if installed.
COMS response time is the most important measure of end-user performance and service delivery. The COMS Interface Agent provides transaction data for each COMS program, program group, user, transaction code (LINC ISPEC), window, and station as defined through the Sightline Install Utility. The data includes Transactions per Second, Average Response Time, Average Queue Depth, CPU, I/O, ReadyQ time, transaction counts, and a response time/transaction count histogram. The statistics reported by COMS can be used to analyze performance problems for on-line transactions, for transaction growth projections, and capacity planning.
Monitoring and managing the performance and resource demands of these mission-critical databases is key to successful and timely processing of most applications. The DMSII Database Interface Agent supports variable monitoring of databases at three different levels: Global metrics only, comprehensive summary of global and all structures, and detail structure monitoring. Key metrics available include memory and buffer usage, overlay statistics, I/O activity and performance, application request statistics, application request delay, controlpoint and syncpoint statistics, audit trail usage and performance metrics, and database status. The DMSII Database Interface Agent provides the data needed for database performance evaluation and tuning, and capacity planning.
The Billing Account and Event Reporting Interface Agent intercepts SUMLOG data and reports resource statistics for “billing accounts,” which can include every job, task, and MCS session run on the Unisys MCP system. The billing account categories can then be used as a basis for a charge-back system. This “billing” data also provides a summary view of resource consumption by “billing account” or workload, and supplements the information provided by Workloads for capacity planning purposes.
This Interface Agent also tracks key events that occur on your system. This can be used to monitor availability of the system, availability of key application services, or any data center or user defined event. The Billing Account and Event Reporting Interface Agent automatically captures system and application outage events. Event records can also be created manually, and any event record can be updated using the Sightline Capacity Power Agent (CPA) utility.
The Event Reporting capability is designed to support improved Data Center management by providing for automatic and manual tracking of system “events” such as halt/loads, memory dumps, job/task/session terminations, and user-defined custom events like application service interruptions on the Unisys MCP system. Events are automatically downloaded to the EA/V workstation for viewing in an EventList window. Sightline CPA can also be used to create tabular reports of existing events. The Event Reporting capability requires both the Sightline Capacity Power Agent, which stores the event records, and the Billing Account and Event Reporting Interface Agent that monitors the Unisys MCP complex for system and user events.
The TCP/IP Interface Agent provides access to TCP/IP network statistics on the Unisys MCP host and for the network interfaces. This set of TCP-specific metrics permits monitoring the host system and network connection activity levels and can be used for capacity planning of the system’s network connections. When coupled with the Sightline Capacity Power Agent facility, long-term network traffic trends can be analyzed.
Disk space usage by user and application can be automatically monitored in “disk accounts” categorized by usercode, disk pack family name, file size and other key parameters. This permits proactive space resource planning and disk space usage chargeback. Sites that use Sightline CPA can also easily perform long-term trend analysis and projection on individual and groups of disk accounts.
The BNAv2 Interface Agent collects BNAv2 network traffic and usage information on Unisys MCP hosts, ICPs, and CP2000 Line Modules. Statistics at the Line and Station level are also available for detailed analysis of specific communications lines and/or station connections. BNA statistics are useful in isolating overloaded lines and network connections, traffic trend analysis and network capacity planning. The BNAv2 Interface Agent also provides information on the most active BNA Connection Groups via EDM’s Event data feature.
This Interface Agent was formerly called the NAP Interface Agent. The change to CAP reflects the Unisys name change from Network Applications Platform to Communication Applications Platform. Users of VoiceSource and other CAP applications can monitor and manage the performance of these telephony applications using the Sightline CAP Interface Agent. This facility provides timely and historical access to a wide variety of important CAP and CAP Application performance metrics including: the Voice Database activity, Voice Database buffer management, overall NAP application-message processing, Network Application (NA) activity, VNMS managed port statistics, and current CAP buffer statistics. CAP Applications are automatically discovered and can be automatically monitored.
Key CAP metrics include: VoiceFile Inuse Pct, Voice Msgs Sent/Sec, Calls Connected/Sec, Transactions below 1 sec response, NAP application Response Time Avg and Queue Depth, Total Ports, Busy Port Percent, Minimum Buffer Cache, and Buffers Allocated/Sec. The CAP Interface Agent also displays up to the top 15 busiest CAP network applications in the Sightline EA/V EventList, providing details on Total Calls/Sec, Incoming Calls/Sec, Outgoing Calls/Sec, Transferred Calls/Sec, Active Dialogs, Call Queue Depth, and Average Response Time. The CAP Interface Agent data enables monitoring call statistics, managing the performance of the CAP Applications, and capacity planning for both the host platform and the telephony interface facilities.