Log Files Monitoring DetailsEach monitored log file is monitored for messages indicative of an error or other important event generated and logged by a server. The log files being monitored are preloaded with some common error messages which are desirable to catch. However, for serious log file monitoring, it is recommended that the user add his own log file patterns that the user would like to be caught. This process is described in detail below.
File IdentificationThis table displays basic information about the log file. The properties in this table are described below:
File StatisticsThis table displays statistical information related to file modification. The properties in this table are described below:
PatternsThis table displays the default patterns used to scan the log file for errors. Patterns may be edited, added, deleted and disabled by accessing the pop-up menu. The columns of this table are described below:
Enabling and Disabling PatternsPatterns listed in the Patterns Table may be enabled and disabled. Enabled patterns are used when scanning the file for matches. Disabled patterns are not used. To disable a pattern:
To enable a pattern:
Adding, Editing and Deleting PatternsUsers may modify which patterns are used when scanning a file by adding, editing and deleting patterns from the table. To add a pattern:
4. Click "OK" to add the pattern. To edit a pattern:
To delete a pattern:
Setting Alarm Limits - The Matches ColumnTo set alarm limits on the number of matches for a given pattern, right-click on the Matches field for that pattern, and select 'Attribute Editor...'. When the Attribute Editor window appears, click on the Alarms tab, and refer to the figure below. Alarm Thresholds: The first six editable fields are alarm thresholds. If a value is specified for an alarm threshold, an alarm will be triggered when the threshold is crossed. There are three types of alarms: the highest priority is the Critical alarm, the next is Alert, and the lowest priority is the Caution alarm. Re-Alarm Every: This field controls how often the module will re-alarm after a given alarm threshold is crossed. For example, consider the situation where a table row has an Alert threshold of greater than 5 matches, and a Critical threshold of greater than 10 matches. Thus, when the number of matches reaches 6, a single Alert alarm will be generated. Now, if Re-Alarm Every is equal to zero, no additional alarms will be generated until the Critical threshold is attained. On the other hand, if Re-Alarm Every is equal to one (the default), a new Alert alarm will be generated each time the number of matches is incremented between 6 and 10. Finally, if Re-Alarm Every is equal to three, a new Alert alarm will be generated each time the number of matches is incremented by three between 6 and 10. In all cases, a new Critical alarm will be generated when the Critical threshold is reached. Alarm Deadband: This value specifies the amount of time that new alarms of the same type are suppressed once an initial alarm of a given type (Critical/Alert/Caution) has been triggered. Note that a new alarm of a different type is not suppressed under any circumstances. Consider an example where the Alarm Deadband is set to 5 seconds: If the Matches field becomes greater than the 'Alert Threshold (>)' alarm limit, the field goes into the Alert alarm state. If the number of Matches for the pattern are incremented again within the the next 5 seconds, a new Alert alarm will not be issued; a new 'Alert Threshold (>)' alarm for this pattern will only be issued if it occurs more than 5 seconds after the previous alarm. However, if the number of Matches has crossed the 'Critical Threshold (>)', a new alarm will be issued, even if this new alarm occurs within 5 seconds of the previous alarm. For almost every log file, it is not necessary to change the Alarm Deadband from 0. In fact, the recommended and default value for the Alarm Deadband is 0. However, when monitoring the log files described below, the Alarm Deadband must be greater than 5 seconds in order to prevent a tight monitoring loop:
Status Message Pattern: This field modifies the status message that is displayed in the Alarms tab when an alarm is generated. The specified custom message is appended to the standard alarm message: <Instance Description> PrimeAlert LogFileMonitor <Description Field in Patterns Table> Pattern Matches In addition to employing generic message text, a number of variable parameters are available for constructing a custom status message:
NOTES:
The default status message pattern is: %param %limit [%value] %message Alarm Window: This field is used to specify the time window within which alarms are triggered . If any alarm thresholds are surpassed at a time outside of this window, an alarm will not be triggered. The default value for this field is blank, specifying that alarms may be triggered at any time.
Resetting the Matches ColumnThe number of matches displayed in the Matches column of the Patterns table can be reset to zero at any time by doing the following:
Note that the 'Last Reset Time' field shows the time at which the Matches column was last reset to zero. The timestamp format is YYYY-MM-DD HH:MM:SS. Setting Alarm Limits - The Matches (Last ...) ColumnTo set alarm limits on the number of matches (in a given time span) for a given pattern, right-click on the 'Matches (Last ...)' field for that pattern, and select 'Attribute Editor...'. When the Attribute Editor window appears, click on the Alarms tab, and refer to the figure below. Alarm Thresholds: The first three editable fields are alarm thresholds. If a value is specified for an alarm threshold, an alarm will be triggered when the threshold is crossed. There are three types of alarms: the highest priority is the Critical alarm, the next is Alert, and the lowest priority is the Caution alarm. Clear Threshold: The Clear Threshold can be used to set a threshold that will clear all alarms when the match-rate drops below the specified value. The clear threshold should be set to NOT overlap with the value of the lowest-severity alarm threshold. For example, if there is a caution threshold of '> 5', then the value inserted into the clear threshold field should be less than or equal to 6. Status Message Pattern: This field has already been described above. The default status message pattern is: > %limit [%value] Alarm Window: This field is used to specify the time window within which alarms are triggered . If any alarm thresholds are surpassed at a time outside of this window, an alarm will not be triggered. The default value for this field is blank, specifying that alarms may be triggered at any time.
Browsing the Log FileTo browse the Log File, the PrimeAlert FileBrowser must be loaded in the same agent as the PrimeAlert for Apache Web Server module. Once loaded, users will be able to launch the FileBrowser from any table within the Logfiles Section. To launch the FileBrowser:
The FileBrowser will automatically load the patterns in the Patterns Table and highlight lines matching any of those patterns.
|