Highlighted
Kallol Absent Member.
Absent Member.
1102 views

Operations Agent PERF Troubleshooting Part 1 [ HP OA 11.1X ]

Introduction to Operations Agent

Kindly refer the attachment 11_X_Componenets.png

11_X_Components.png

 

HP Operations agent 11 combines the features of the HP Operations agent, HP Performance Agent and HP GlancePlus together under one installer.

The HP Operations agent helps you monitor a system by collecting metrics that indicate the health, performance, and availability of essential elements of the system. While HP Operations Manager (HPOM) presents you with the framework to monitor and manage multiple systems through a single, interactive console, the HP Operations agent deployed on individual nodes helps you gather vital information to facilitate the monitoring process.

When you use the HP Operations agent in conjunction with HPOM and Smart Plug-ins (SPIs), you can add the capability to monitor business applications running on monitored systems. When used in isolation (in an environment where HPOM is not deployed), you can use the data store of the HP Operations agent to read system performance data. Based on your requirement, you can install the HP Operations agent in one of the following uses:

  • Centralized monitoring with HPOM: In this scenario, HPOM and the HP Operations agent together build up a distributed monitoring environment that helps you manage heterogeneous systems and applications.
  • Performance monitoring of an isolated system: When installed without the presence of HPOM, the HP Operations agent enables you to monitor the health and performance.

Operations agent is supported on a wide variety of operating system platforms, high-availability environments and virtualization platforms.

HP Operations agent includes the following major operational components:

  • Shared Components (aka L-Core components)
  • Operations Monitoring Component (aka Event/Action component)
  • Performance Collection Component
  • Real-Time Metric Access Component (aka perfd)
  • Real-Time Measurement Component (aka RTM)
  • Glance

Operations agent ideally is combination of Monitoring and Collection components.

Operations Monitoring Components

The Operations Monitoring Component (aka the Event/Action component) builds up the monitoring and messaging functionality of the HP Operations agent. It can perform the following tasks:

  • Responsible for scheduling and running policies deployed from the management server
  • Generate alert messages when the monitored object violate preset thresholds or based on conditions configured in the policies
  • Forward the messages to the HPOM console
  • Execute actions (scheduled actions, automatic or operator-initiated actions based on policy conditions)

Performance Collection Components

The Performance Collection Component (aka the Performance Agent) builds up the metric collection and store functionality of the HP Operations agent. It can perform the following tasks:

  • Collection of System Performance/User defined and application Metrics
  • Store the collected data in a local store for near real-time analytics
  • Provide the collected data at real time to interested consumers locally (Glance) or remotely (via RTM to PM). Currently only system performance data is provided.
  • Provide stored/historical data for local monitoring (opcmona/perfalarm) or graphing/reporting to other HP Products (PM, SHR).

Performance Collection

Kindly refer the attachment Performance_Agent.png

Performance_Agent.png

The main features of various components of the Performance Collection Component are:

Performance Component (scope)

  • Collect, timestamp and log the performance metrics in log files. The scope log files contains the performance metrics collected by the scope collector
  • Parm file is the main configuration file which controls the collection of metrics such as the thresholds, classes and collection interval
  • This also contains the application definitions
  • The LIBSCOPE library contains all the methods to read the SCOPE and DSI log files
  • It used the NUMS library to collect performance data

Performance Alarming Component (perfalarm)

  • It is a real time alarming component which sends alerts to the Operations Management Console based on the alarm thresholds defined in the ALARMDEF files

Embedded Performance Component (coda)

  • It is the main component which supplies the data logged by SCOPE to the external entities where the collected data is used for the maximum.
  • Ddflog is used to log Data into Coda data store. SPIs use this interface (directly or indirectly using dsi2ddf) to store data in the Coda data store.
  • It uses the https communication to supply the data contained in the scope log files and the Coda data store to its clients such as PM, OA, SPI, BSM, PERFALARM etc.

Glance

  • Display the metrics in real time on X-Windows based GUI or on the curses based GUI on the terminal

Real-time Collection Component (perfd)

  • The main functionality of perfd is to provide real time metrics to RTMD for updating the Performance Manager live graphs.
  • PERFD.INI contains the parameters to configure the real time collection component PERFD Real-time Publish/Subscribe (rtmd)
  • RTMD is a module which helps the Performance Manager to get the real time metrics collected by PERFD via BBC sockets

Measurement Interface Component (midaemon)

  • It is a specialized daemon program which is running with the highest priority to provide the Performance Metrics and Transaction Metrics to the NUMS library
  • On HP-UX the OS is specially instrumented for sharing information with midaemon to build the performance metrics
  • Shared memory is the way of communication between the midaemon and NUMS library
  • On Platforms other than HP-UX this only provides the Transaction Metrics to NUMS library

Transaction Tracking Daemon (ttd)

  • This program supplies the ARM instrumented applications with the required Application and Transaction IDs for tracking the transactions and response times
  • The configuration file called ttd.conf contains the SLO (Service Level Objectives) and configurations for Transaction logging

Virtualization Collectors (viserver/lsdaemon)

  • These are extensions of the NUMS Library, which collects the metrics data on special environments where the collection requires special mechanisms
  • VIServer is a Java based program which helps the NUMS library to collect data from remote ESX Servers and shares it with NUMS Library via TCP/IP Sockets
  • LSDAEMON is an external program which collects data on AIX LPARs and shares it with NUMS Library via shared memory

Now that we have fair knowledge of OA 11.1X, lets dive deeper into the next part for troubleshooting different category of issues and individual components.

Operations-Agent-PERF-Troubleshooting-Part-2-HP-OA-11-1X

Thanks and Regards,
_______________________________________________________
Kallol Roy

"Be kinder than necessary, for everyone you meet is fighting some kind of battle."
The opinions expressed above are the personal opinions of the authors, not of Micro Focus. By using this site, you accept the Terms of Use and Rules of Participation. Certain versions of content ("Material") accessible here may contain branding from Hewlett-Packard Company (now HP Inc.) and Hewlett Packard Enterprise Company. As of September 1, 2017, the Material is now offered by Micro Focus, a separately owned and operated company. Any reference to the HP and Hewlett Packard Enterprise/HPE marks is historical in nature, and the HP and Hewlett Packard Enterprise/HPE marks are the property of their respective owners.