Documentation for up.time 7.0 If you are looking for the latest version view

Unknown macro: {spacejump}

or visit the current documentation home.

Accessing and Exiting up.time

Before logging into up.time , you will need a user name and password from your system administrator. Your system administrator will provide assistance if this is your first time logging into the application.

Setting Up the Administrator Account

The first user to log into up.time should be the system administrator. While the administrator account has the default user name admin , you will have to set the password and email address for the administrator account. You will only need to do this the first time that you log into up.time .

To set up the administrator account, do the following:

  1. Enter the following in the address bar of a Web browser:
    http://<uptime_hostname>:<port>
    Where <uptime_hostname> is the name or IP address of the server that is hosting the Enterprise Monitoring Station. For example:
    http://localhost:9999
    The up.time log in window opens in a Web browser.
  2. Enter the password for the administrator in the Password field.
  3. Re-enter the password in the Confirm Password field.
  4. Enter your email address in the Administrator’s Email field.
  5. Click the Login button.

Accessing up.time

Once an administrator sets up your up.time account, you can navigate and log in to the Enterprise Monitoring Station.

To start up.time , do the following:

  1. Start a Web browser.
  2. Enter the following in the address bar of the Web browser:
    http://<uptime_hostname>:<port>
    Where <uptime_hostname> is the name or IP address of the server that is hosting the Enterprise Monitoring Station.
    The up.time log in window opens in the Web browser.
  3. Enter your assigned user name in the User Name field.
  4. Enter your assigned password in the Password field.
  5. Click the Login button.

Exiting up.time

To exit up.time , click the Logout button in the top right corner of the screen.

Viewing System and Service Information

You can view information about the following:

  • basic configuration of systems in your environment
  • services and service groups assigned to the system
  • user groups assigned to the system

Viewing System Information

To view system information, do the following:

  1. In the Global Scan or My Infrastructure panels, click the name of a system.
    The general information for the system appears in the sub panel.
  2. Click the Infotab, and then click one of the following options in the Tree panel:
    • Info & Rescan
      Lists the basic information about the system, including the following:
      • the display name of the system in up.time
      • the host name
      • the number of processes the monitors will retrieve
      • whether or not the system is being monitored
      • the name of the domain on which the system resides (e.g., uptimesoftware.com )
      • the name and version of the operating system that is running on the system
      • the number of CPUs on the system
      • the amount of memory, in megabytes, on the system
      • the size of the paging file, in megabytes, on the system
      Click the Rescan Configuration button to refresh the configuration information for an agent or a Net-SNMP host. You would do this, for example, if a disk was added to the system. A progress window appears.
      When the message Configuration Rescanning Completed appears, click Close Window . Information about the configuration changes, if any, appears in Configuration Changes section of the subpanel.
      Note - If the system that you selected in step 1 is a node, then only the following information appears: the display name and host name of the node, its parent group, and whether or not the node is monitored.
    • CPU Information
      Lists the speed (in MHz) of all of the CPUs on the system.
    • Network
      Lists the network interfaces on the system, as well as the IP addresses of those interfaces.
    • Disks/File System
      Lists the disks that are on Solaris and Linux systems and the names of the file systems that up.time is monitoring.
    • Poll Agent
      Displays the output from an up.time agent that you suspect may have a problem. You can forward the output to uptime software Client Care when you encounter problems with up.time .
    • Services
      Lists the services assigned to the system, as well as the interval (in minutes) at which the services are checked.
    • User Groups
      Lists the user groups that are associated with the system.

Viewing Service Information

To view system information, do the following:

  1. In the Global Scan or My Infrastructure panels, click the name of a system.
  2. Click the Services tab in the Tree panel.
  3. Click one of the following options in the Tree panel:
    • Status
      Lists the status of each service assigned to the system, for example:
      up.time agent running on subway [up.time agent running on subway, up.time agent 4.0 solaris
      Note -  An arrow at the end of a status message indicates that there is more text. Hold your mouse over the arrow to view the full message.
      When up.time issues an alert, you can acknowledge the alert in the Status subpanel. For more information, see Acknowledging Alerts.
    • Trends
      Displays one or more graphs that chart the status of the services associated with a host. For more information on what each status means, see Understanding the Status of Services.
    • Outages
      Lists, in tabular format, the services that have suffered outages, along with the time at which the outage occurred. The Outages table also lists all changes to the states and substates for services and host checks - for example, from OK to CRIT and then from CRIT to OK .
      As well, up.time displays a message describing the outage - for example:
      Socket error has occurred connecting to elinux 
      Error text: Connection timed out: connect
      If you are using the Splunk IT search engine with up.time , the Splunk icon appears beside the names of services that are in WARN or CRIT states. You can click the icon to check the Splunk logs for information about the outage.
    • Availability
      Lists the state - OK , WARN , CRIT , MAINT , UNKNOWN - of the monitors that are associated with a specific host or device, as well as:
      • the amount of time that the services have been in each state and the total of all times
      • the percentage of time each service has been in each state
        Optionally, click the Generate Graph button to display pie charts that graph the status of each service.
    • Manage Services
      Lists the following information about the services associated with a particular host:
      • the name of the service
      • the service group, if any, to which the service belongs
      • the monitors, if any, associated with the service
        If the host is part of a service group, the services for all of the hosts that are members of the group appear in the Manage Services subpanel.
        Click the name of the service to view information about that service. You can edit the service information, as well as the Alert Profiles and Action Profiles associated with the service by clicking the appropriate button in the subpanel.
        You can add services instances by clicking the Add Service tab in the Manage Services subpanel. The services that you add do not appear in the Manage Services but in the Service Instances subpanel. For more information about adding service instances, see Using Service Monitors.
    • Host CheckList the basic checks (for example, a ping) for a system.
    • Maintenance
      Lists whether or not there are any maintenance periods scheduled for the system. For more information on maintenance periods, see Scheduling Maintenance.
  4. Optionally, click Service Metrics to generate a graph that visualizes retained data over a given period of time. For more information about retained data, see Understanding Retained Data.
  5. To generate a graph, do the following:
    • Select the date range for the graph from the Date Range area. For more information, see Understanding Dates and Times.
    • In the Current Retained Service Metrics area, select the retained data variables that you want to graph
  6. Click Generate Graph

Searching and Filtering

If you have a large number of hosts on your system, you can use the search and filtering functions in the up.time Web interface to quickly display and view information about specific hosts.

Using the Search Box

You can use the search box at the top of the up.time Web interface to display the basic information about a particular host.

To use the search box, do the following:

  1. From anywhere in the up.time Web interface, enter any of the following information in the Search box:
    • The name of the system for which you want to search.
      Note - You can enter a partial name in the Search box. For example, if you want to display all systems whose names start with Web , enter Web in the Search box.
      Details about the architecture of the servers. For example, to use an operating system as the search criteria enter Linux in this field.
    • Any information that may appear in the Custom fields in the profile for the system.
  2. Click Go.
    • The following information is displayed:
    • name of the host
    • description of the host (if any)
    • the operating system and type of hardware on which the host is running
    • any information in the four custom fields in the system profile (e.g., the job being done by the system, and its physical location)
      For more information, see Editing a System Profile.

Filtering Service Instances

If you have a large number of hosts and want to view information about a particular service instance associated with those hosts, you can filter out the services that you do not want to see in the Service Instances subpanel.

To filter service instances, do the following:

  1. On the up.time tool bar, click Services .
  2. In the Tree panel, click View Service Instances .
  3. Enter text in one of the following fields in subpanel:
    • Name
      The name of a particular service instance, for example PING-Server1 .
      You can enter partial names of service instances in this field. For example, if you want to filter on instances that contain the text Mailbox , enter Mailbox in the field.
    • Host
      The name of a host with which the service is associated. This can be the actual name of the host or the display name in the up.time Web interface.
    • Monitor
      The name of a particular monitor on which you want to filter. For example, Ping or LDAP .
      You can enter partial names of monitors in this field. For example, if you want to filter on File System Capacity , enter Capacity in the field.
  4. Click Filter By .
    All service instances that you have permissions to view and that match the filtering criteria appear in the subpanel. If, for example, only 12 of the service instances match your criteria, a message like the following one appears in the subpanel:
    Search found 12 out of 21 services
  5. To view all matches, click the Show All button.
  6. To remove the filter criteria and restore the complete list of services, click Clear .

Audit Logging

up.time can record changes to the application's configuration in an audit log. The details of the configuration changes are saved in the file audit.log , found in the logs directory.

Note - Windows Vista users can find the audit log in the Virtual Store instead of the default location
(i.e., C:\Users\uptime\AppData\Local\VirtualStore\
Program Files\<uptime-install-directory>

There are many uses for the audit log. For example, you can use the audit log track changes to your up.time environment for compliance with your security or local policies. You can also use the audit log to debug problems that may have been introduced into your up.time installation by a specific configuration change; the audit log enables you to determine who made the change and when it took effect.

The following is an example of an audit log entry:

2006-02-23 12:28:20,082 - dchiang: ADDSYSTEM [cfgcheck=true, port=9998, number=1, use-ssl=false, systemType=1, hostname=10.1.1.241, displayName=MailMain, systemSystemGroup=1, serviceGroup=, description=, systemSubtype=1]

Enabling the Audit Log

By default, the audit log is disabled. To enable it, edit the uptime.conf file, which is located at the root of the up.time installation directory:

  • /opt/uptime on Solaris
  • /usr/local/uptime/ on Red Hat and SLES
  • C:\Program Files\uptime software\uptime on Windows

In the uptime.conf file, locate the “ auditEnabled= ” entry and modify it to be “ auditEnabled=yes ”. If the entry does not exist, add the entry to the file.

An arrow at the end of a status message indicates that there is more text. Hold your mouse over the arrow to view the full message.

  • No labels