Using the Monitoring Tools
In addition to viewing server status from the Server Status page, you can click the Monitoring Tools link to display a page that has links to additional monitoring tools. These links provide a convenient way to get information about how your system is running from your browser instead of bringing up another program.
|
|
To use the monitoring tools, log on as a member of the site Administrators group or as a user who is listed in the wt.util.jmx.AccessUtil.siteAdministrators property. For details on using this property, see “Setting Up Your JMX Environment” in the Windchill Help Center.
|
The following list describes how to use each of the links on the System Health Monitoring Tools page:
• Clicking the Server Status Page link returns you to the Server Status page.
• Clicking the Windchill Cache Statistics link presents a page that initially lists each cache and details about the cache usage. The statistics include where the cache resides, the number of entries and capacity of the cache, and the number of hits and misses.
To limit the number of caches shown in the list, you can fill in values in the fields above the list and click OK. For example, entering 42 in the Minimum Misses field, limits the list to those caches with 42 or more misses.
• Clicking the Performance Feedback Settings link presents a page from which you can adjust log levels, data output, and the formatting of several key performance loggers. Additionally, you can enable and disable a performance footer.
• Clicking the Log Levels link presents a page from which you can display or set log levels for server manager, method server, or servlet engine loggers for current server processes. First select one or more processes and then browse for the logger.
• Clicking the Log Comment link presents a page where you can enter a comment that is then stored in server manager and method server logs. If your environment is set up as a cluster, the logs include logs stored on all nodes in a cluster.
Use this link to add log entries that can do things like demarcating the start and end of a load test or indicating when some configuration fix (such as a file permission fix) was made that did not require a server restart.
Essentially, this link allows an administrator to add a comment to the server logs. Consider adding comments when an important, known state change has occurred that the system itself would not detect and the addition of a comment in the logs could ease later troubleshooting or analysis.
• Clicking the Log File Viewer link presents a page from which you can search, list, and view server log files.
• Clicking the Persisted Log Events link presents a page from which you can browse recently persisted log events and view event details. By default, the Minimum Severity Level is set to WARN and the age of the listed events is from the last six hours. The level selected is the lowest level of events returned; therefore selecting WARN returns warning, error, and fatal events. Selecting OFF returns only log events whose level is OFF, which normally do not exist.
• Clicking the Log Event Histogram link presents a page from which you can decide on the age of the log events to display and select the Minimum Severity Level to include. Available error levels are listed in the order of severity, lowest (OFF) to highest (All). For example, using the default (WARN) level includes log events associated with the FATAL, ERROR, and WARN levels.
Clicking OK presents a histogram in tabular form that shows the list of events by logger name, level, and count.
• Clicking the Method Context and Servlet Request Samples link presents a page from which you can browse performance sampling data for active method contexts and servlet requests. If no times are specified, the default time interval is the last fifteen minutes.
To limit the results to a specific context or request, enter its ID in the Context/Request Id of Interest field.
• Clicking the Cluster-wide Stack Traces link presents the <InService>/wtcore/jsp/jmx/clusterStacks.jsp page. This page presents stack traces that are taken at the time you click the link. The page is divided into the following sections:
◦ Server Manager Stack Traces
This section has stack trace information for each server manager in the cluster. A server manager is identified by its <process_ID>@<host_name>, and the date and time of the stack trace is included. For example, assume:
▪ The process ID is 8356.
▪ The host name is host123.
▪ The date and time is 19th of December 2011 at 14:50:42.513 CST.
Then, the following lines identify the beginning of the stack trace:
Server Manager: 8356@HOST123
----------------------------
Full thread stack trace (8356@HOST123; 2011-12-19 14:50:42.513 -0600)
◦ Method Server Stack Traces
For InService, this section duplicates the server manager stack trace information.
|
|
The time designations are different in the previous examples because the JVM’s default time zone is used. In this case, the server manager JVM was set to CST and the method server JVM was set to GMT. The actual time difference between the two stack traces was only 0.125 seconds.
|
• Clicking the Top SQL Sample Intervals link presents a page from which you can view information on the SQL statement executions that took the longest total time within a given time interval.
The defaults are a time interval of the last six hours and a search for SQL statements in all JVMs. Entering a JVM limits the results to SQL statements executed in that JVM.
Clicking OK returns a table summarizing the samples found. Each table row includes a link to details about the sample.
• Clicking the Java Process Information link presents a page from which you can view persisted status information for historic and active server processes.
Use this page to search for processes. For example you can locate information about a method server that recently died by clicking OK to display a list of recent processes.
If you do not enter a time interval, the default interval is the last three days.
To limit the search to a specific host, you can enter the following string in the JVM field:
*@<hostname>
After you locate a process from the list displayed, then click the corresponding link to display persisted information about the process, such as the JVM arguments for the process and various charts containing information similar to what is seen on the Server Status page for active processes.
• Clicking the Client User-Agent Usage link presents a table that lists the users who accessed the system during a period of time and the corresponding agent that was used. The agents include browser versions and Java plug ins.
• Clicking the Export System Health/Performance Tables to Client link presents a page where you enter the starting and ending times for which you want the export to cover. Use this link when you want to provide PTC Technical Support with data from your database but you cannot send the data directly from your system. Instead, export the data to a client file on your system and then transfer that file to a location from which you can send it to technical support. If you want to check the exported file content, you can import it by opening a Windchill shell and using java wt.util.jmx.ImportPerfTables.
The capability to export these tables is primarily intended to facilitate interaction with technical support by exporting this data so that it can be sent to technical support personnel. They can then import the data into one of their systems and analyze it without having access to your system.
• Clicking the Export System Health/Performance Tables to Support link presents a page where you enter the starting and ending times for which you want the export to cover. Use this link when you want to provide PTC Technical Support with data from your database.
Before doing this export, you must have entered a valid PTC Service Contract Number (SCN) and a valid support call number. For details on setting the SCN, see
Viewing and Sending Server Information.
The capability to export these tables is primarily intended to facilitate interaction with technical support by allowing this data to be automatically sent to them. They can then import the data into one of their systems and analyze it without having access to your system.
