IBM Tivoli Monitoring > Version 6.3 Fix Pack 2 > Installation Guides > Agent Installation Guides > UNIX Agent Installation Guide > Overview of the agent

IBM Tivoli Monitoring, Version 6.3 Fix Pack 2

---

New in this release

For version 6.3 Fix Pack 2 of the monitoring agent, enhancements include:

• The Tivoli Common Reporting data model exposes the Managed System List. You can use the Managed System List in combination with, or as an alternative to, the Managed System Name. This capability is available for custom reporting only and the specified metrics are aggregated using the default aggregation function.

• The Infrastructure Management Dashboards for Servers (Server Dashboards) managed system dashboard has new tabs: Properties, CPU, Memory, Disk, and Network. Several tabs have a new time selector bar for changing from real time to showing historical data; and the situation event results dashboard Details tab has a new time selector bar for setting a time range of data samples before or after the event time.

• The monitoring agent complies with the Federal Information Processing Standard (FIPS) 140-2. This computer security standard requires stronger checksum algorithms (for example, SHA-256 and SHA-512) when you define situations for checking file changes.

• An internal caching mechanism improves agent performance, in terms of response time and CPU consumption while collecting process information. The agent updates process information in cache, related to process PID, command, and arguments, every 120 seconds by default. To change the cache refresh time from this default value, specify the cache refresh value for the environment variable KUX_PROCESS_CMD_SAMPLE_SECS (minimum valid value is 30 seconds). If the environment variable is set to 0, the internal caching mechanism is disabled.

• Several reports were added to Tivoli Common Reporting, including the Top n Process Usage by WPAR report, WPARs Configuration report and WPAR Utilization report.

• Various metrics were ported from the AIX Premium agent to the Monitoring Agent for UNIX OS.

  • New attribute groups include AIX MPIO Attributes, AIX MPIO Status, AIX Network Adapters, and AIX System IO. To customize the sampling interval for AIX System IO metrics, specify the value of the KUX_MAP_SAMPLING_INTERVAL environment variable (default value: 60 seconds). To specify this sampling interval as variable, set 0 for the environment variable. As a result, the data sampling occurs when an agent receives the request.

  • The AIX Physical Volumes attributes group includes Number of Stale Partitions.

  • The Process attributes group includes Text Size.

  • The Disk Performance attributes group includes Avg Read Transfer MS, Avg Write Transfer MS, Failed Read per Sec, Failed Writes per Sec, Max Read Service MS, Max Request In WaitQ MS, Max Write Service MS, Min Read Service MS, Min Request In WaitQ MS, Min Write Service MS, Read Timeouts per Sec, and Write Timeout per Sec.

  • The Network attribute group includes Domain, Gateway, and Mask.

  • The AIX LPAR attributes group includes CPU Capacity Increment, Max Dispatch Latency, Min Req Virt CPU, Min Virt CPUs, and Num Hypervisor Calls per Sec. The Num Hypervisor Calls per Sec attribute is collected using the perfstat_hyperstat_total() system API, supported by AIX 6.1 TL5 FP2 or later.

  • New situations associated with the Disk Usage node, Network node, and Process node.

  • The AIX MPIO Storage Information workspace has views that show the AIX Multi-Path I/O (MPIO) Attributes, AIX Connection Status, and AIX Storage Devices Utilization on the current LPAR.

  • The AIX Network Adapters workspace displays data related to utilization and errors per network adapter.

• The AIX Physical Volumes attributes group now includes the Number of Stale Partitions attribute.

• The UNIX Memory attributes group now includes Available File Cache MB (AIX), Computational Memory MB (AIX), and Non Computational Memory MB (AIX). The System Memory view in the System Details workspace reports these new attributes. In addition, the UNIX Memory attributes group now includes Percent Real Memory Process (AIX), Percent Real Memory System (AIX), Percent Page Replacement Memory Current Value (AIX), Percent Page Replacement Memory Min Value (AIX), and Percent Page Replacement Memory Max Value (AIX).

• The AIX Memory Per Page attributes group contains information about memory statistics per page size. The AIX Memory Details workspace contains views of AIX-specific data collected for the Unix Memory group and the AIX Memory Per Page group. To customize the sampling interval for Zero Frames Per Sec (MB), Page Steals Per Second (MB), Paged In Pages from Page Space Per Sec (MB), Paged Out Pages from Page Space Per Sec (MB), and Page Scans Frames By Clock Per Sec (MB) metrics, specify the value of the KUX_MAP_SAMPLING_INTERVAL environment variable (default value: 60 seconds). To specify this sampling interval as variable, set 0 for the environment variable. As a result, the data sampling occurs when an agent receives the request.

• The Disk Performance attributes group now includes Volume Group Name (AIX). The Disk Performance view in the Disk Usage Details workspace reports this new attribute.

• For the Utilization Details for Single Resource report, you can specify the resources to display (CPU, Memory, Disk, Network, or Process).

• In addition to monitoring the status of the mount_stat, aixdp_daemon, and stat_daemon subprocesses used by the UNIX OS Agent to collect data from the system, you can monitor the health of the stat_daemon children: kux_vmstat for CPU and memory statistics, ifstat for network interface statistics, nfs_stat for NFS and RPC statistics, and kuxdstat (or iostat, for AIX) for disk I/O statistics. You can disable the kuxdstat data provider at startup, by specifying the environment variable KUX_DISABLE_UNIXDPERF=TRUE. As a result, the status of the Data Provider kuxdstat (or iostat for AIX) is set to "Disabled". The Data Collection Status attribute group now includes Process ID.

• A new situation, UNIX_Agent_Processes_Failure, is associated with the System Information node.

For version 6.3 of the monitoring agent, enhancements include:

• Various metrics were ported from the AIX Premium agent to the Monitoring Agent for UNIX OS.

  • New attribute groups include AIX Logical Volumes, AIX Physical Volumes, AIX Volume Groups, Top CPU Processes, Top Memory Processes, and UNIX Devices.

  • The UNIX workspace, Process workspace, and All Processes workspace were updated with revised views to incorporate data that is offered by the Top CPU Processes, Top Memory Processes, and UNIX Devices attribute groups.

  • The AIX Storage workspace contains views of data that is related to logical volumes, physical volumes, and volume groups. The views for this workspace include the Physical Volume Sizes bar chart, Physical Volume Details table view, Volume Group Sizes bar chart, Volume Group Details table view, Logical Volume Sizes bar chart, and Logical Volume Details table view.

• The AIX Devices Status workspace was superseded by the Devices Status workspace. In addition, the UNIX_Device_Stopped_Warning situation indicates whether a specific UNIX device stopped.

• The Data Collection Status attributes group reports on the health of internal data collectors of the Monitoring Agent for UNIX OS. The Data Collection Status table view of the UNIX workspace provides specific details.

• The UNIX Memory attributes group now includes Percent Available File Cache (AIX), Percent Computational Memory (AIX), and Percent Non Computational Memory (AIX). The System Virtual Memory view in the System Details workspace reports these new attributes.

• For attribute values calculated as an average of the cumulative CPU ticks between two samples, the sample time differs depending on how the agent is invoked to return the values. If the agent is invoked to return the values on-demand (for example, after a workspace refresh), the default sample time is 30 seconds for total CPU metrics and 60 seconds for the CPU metrics per process. If, however, the agent is invoked to return the values by a situation or a historical collection request, the sample time is the same as that of the situation or of the collection interval. The affected attributes include:

  • SMP CPU attribute group: User CPU, System CPU, Idle CPU, Wait I/O, CPU Busy, and CPU Usage attributes

  • SMP CPU attribute group, for SUN Solaris OS agents: Minor Faults, Major Faults, Cross Calls, Interrupts, Interrupts As Threads, Context Switches, Involuntary Context Switches, Thread Migrations, Spins On Mutexes, Spins On RW Locks, and System Calls attributes

  • Process attribute group: CPU Pct attribute

  • Top CPU Processes attribute group: CPU Pct attribute

  • Top Memory Processes attribute group: CPU Pct attribute

  You can customize the sampling intervals by specifying two variables in the ux.ini file: KUX_CPUSTAT_SAMPLE_SECS for the total CPU metrics (default value: 30 seconds), and KUX_PROCESS_SAMPLE_SECS for the CPU metrics per process (default value: 60 seconds). If these variables are set to 0, the sampling interval is variable: the samples are taken when the requests come to the agent (for example, at each workspace refresh), and the sampling interval is the difference in time between last two samples (with a minimum of 5 seconds).

  The CPU statistics measurements are provided by system API. Therefore, the KUX_IGNORE_MPSTAT, KBB_HPUX_SAR, and KBB_HPUX_VMSTAT environment variables are no longer required. Even if the variables are specified, they are ignored.

• The Summarization and Pruning agent automatically creates and maintains the shared dimensions tables. For instructions to enable this feature, see "Configuring the Summarization and Pruning agent to maintain the dimension tables" in the IBM Tivoli Monitoring Administrator's Guide. To enhance this feature for the OS Agents Reports package, the installer now prompts you to provide JDBC connection details and credentials for the Tivoli Data Warehouse database. This RegisterPackage script execution step inserts data into the WAREHOUSETCRCONTROL table. After this step, the MANAGEDSYSTEM table and the TIME_DIMENSION table are kept up to date automatically by the Summarization and Pruning agent. However, if you opt not to use this feature and prefer, instead, to manually maintain the dimensions tables, skip this step. For instructions to perform any required manual steps, see "Manually creating and maintaining the dimension tables" in the IBM Tivoli Monitoring Administrator's Guide.

• The agent provides ComputerSystem and IPAddress resources for the Open Services for Lifecycle Collaboration Performance Monitoring (OSLC-PM) service provider. The service provider registers monitoring resources with the Registry Services. Registry Services is a Jazz for Service Management integration service that provides a shared data repository for products in an integrated service management environment.

• The IBM Tivoli Monitoring Infrastructure Management Dashboards for Servers is a web-based application that runs in the Dashboard Application Services Hub. The server dashboards give the overall status of the service areas in your managed network. Use the server dashboards to assess the event and system status of your managed network that is filtered by your area of responsibility. The information ranges from a high-level overview of all managed system groups and the situation events that are associated with them, to more detailed dashboards with key performance information about the selected group, managed system, or situation event.

Parent topic:

Overview of the agent

---

+

Search Tips   |   Advanced Search