| Oracle® Enterprise Manager Framework, Host, and Third-Party Metric Reference Manual 10g Release 2 (10.2) Part Number B16230-01 |
|
|
View PDF |
| Oracle® Enterprise Manager Framework, Host, and Third-Party Metric Reference Manual 10g Release 2 (10.2) Part Number B16230-01 |
|
|
View PDF |
The oracle_emd target is a representation of the Oracle Management Agent. The Oracle Management Agent is the Management Agent used by Oracle Enterprise Manager. This target type exposes useful information required to monitor the performance of the Management Agent.
Most of the help topics in this helpset use the term Management Agent to refer to the Oracle Management Agent.
The EMD Process Statistics provides information about the performance and resource consumption of the Management Agent process. This metric is collected by default on an interval of 1038 seconds. A value that can be changed in the default collection for the oracle_emd target.
The amount of resident memory used by the agent and all of its child processes in KB.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 1038 Seconds |
Data Source
Varies. On UNIX platforms this is derived from the ps command.
User Action
The default warning and critical threshold values for this metric are set higher than what is expected to be necessary in many cases. You may give a lesser value for the warning and critical thresholds based on the number and types of targets that are being monitored by the Management Agent.
The amount of virtual memory used by the agent and all of its child processes in KB.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 1038 Seconds |
Data Source
Varies. On UNIX platforms this is derived from the ps command.
User Action
The default warning and critical threshold values for this metric are set higher than what is expected to be necessary in many cases. You may give a lesser value for the warning and critical thresholds based on the number and types of targets that are being monitored by the Management Agent.
The CPU Usage metric provides the CPU consumption as a percentage of CPU time at any given moment in time. The number is a summation of the CPU consumption of the Management Agent process and any of its child processes (and their child processes and so on). Child processes are sometimes created by the Management Agent in the course of evaluating a metric or running a job.
Metric Summary
The following table shows how often the metric's value is collected and compared against the default thresholds. The 'Consecutive Number of Occurrences Preceding Notification' column indicates the consecutive number of times the comparison against thresholds should hold TRUE before an alert is generated.
Table 1-1 Metric Summary Table
| Target Version | Evaluation and Collection Frequency | Upload Frequency | Operator | Default Warning Threshold | Default Critical Threshold | Consecutive Number of Occurrences Preceding Notification | Alert Text |
|---|---|---|---|---|---|---|---|
|
All Versions |
Every 1038 Seconds |
After Every Sample |
> |
10 |
20 |
4 |
Agent CPU consumption is %value%%% |
Data Source
The source for this metric is the UNIX ps command.
User Action
A large CPU consumption will cause the entire system to slow down. The cause could be the Management Agent process itself or any of its child processes. To analyze what is causing the problem, use the Solaris "top" system command and look out for any Perl or Java processes that seem to be consuming excessive CPU (%).
This metric records the number of files currently opened by the Management Agent process. The file types that constitute this number are: regular files, links, sockets, directories and name pipes.
Metric Summary
The following table shows how often the metric's value is collected and compared against the default thresholds. The 'Consecutive Number of Occurrences Preceding Notification' column indicates the consecutive number of times the comparison against thresholds should hold TRUE before an alert is generated.
Table 1-2 Metric Summary Table
| Target Version | Evaluation and Collection Frequency | Upload Frequency | Operator | Default Warning Threshold | Default Critical Threshold | Consecutive Number of Occurrences Preceding Notification | Alert Text |
|---|---|---|---|---|---|---|---|
|
All Versions |
Every 1038 Seconds |
After Every Sample |
> |
800 |
900 |
2 |
Number files opened by Agent is %value% |
Data Source
The source of this information is the UNIX pfiles command. On non-UNIX platforms this will not be collected. On Windows platforms, refer to the File Handles Open metric.
This metric records the number of file handles currently opened by the Management Agent process.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 1038 Hours |
Data Source
This metric is collected on Windows platforms and is not collected on UNIX platforms. For UNIX, use the "Number Files Open" instead. It is gathered by an agent api.
This metric shows the number of threads currently created by the Management Agent process.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 1038 Seconds |
Data Source
The source of this metric is the UNIX ps command.
The process ID is the process ID of the Management Agent.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 1038 Seconds |
Data Source
The source of this is the Perl getppid function.
The Resident Memory Utilization is the physical memory usage as a percentage of total memory available.
Metric Summary
The following table shows how often the metric's value is collected and compared against the default thresholds. The 'Consecutive Number of Occurrences Preceding Notification' column indicates the consecutive number of times the comparison against thresholds should hold TRUE before an alert is generated.
Table 1-3 Metric Summary Table
| Target Version | Evaluation and Collection Frequency | Upload Frequency | Operator | Default Warning Threshold | Default Critical Threshold | Consecutive Number of Occurrences Preceding Notification | Alert Text |
|---|---|---|---|---|---|---|---|
|
All Versions |
Every 1038 Seconds |
After Every Sample |
> |
20 |
30 |
1 |
Agent resident memory utilization is %value%%% |
Data Source
The source of this information is the UNIX ps system command.
This metric represents the amount of physical memory usage by the Management Agent process and all of its child processes in KB.
Metric Summary
The following table shows how often the metric's value is collected and compared against the default thresholds. The 'Consecutive Number of Occurrences Preceding Notification' column indicates the consecutive number of times the comparison against thresholds should hold TRUE before an alert is generated.
Table 1-4 Metric Summary Table
| Target Version | Evaluation and Collection Frequency | Upload Frequency | Operator | Default Warning Threshold | Default Critical Threshold | Consecutive Number of Occurrences Preceding Notification | Alert Text |
|---|---|---|---|---|---|---|---|
|
All Versions |
Every 1038 Seconds |
After Every Sample |
> |
128000 |
256000 |
1 |
Agent resident memory utilization in KB is %value% |
Data Source
The source of this data is the UNIX ps system command.
User Action
The default warning and critical threshold values for this metric are set higher than what is expected to be necessary in many cases. You will probably want to lower the warning and critical thresholds to values that work well for the number and types of targets that are being monitored by the Management Agent.
The Virtual Memory Utilization (VMU) metric provides a sum of the VMU usage of the Management Agent and all of its child processes (and their child processes and so on). Child processes are sometimes created by the Management Agent in the course of evaluating a metric or running a job.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 1038 Seconds |
Data Source
The source of this information is the UNIX ps system command.
User Action
Large virtual memory utilization will also slow the system down. On UNIX machines, use the "top" command to see what processes are consuming this memory. Look out for Perl and Java processes as well as the obvious emdaemon process (the Management Agent process itself.)
Virtual memory utilization growth (%) shows the percentage growth of the virtual memory percentage usage of the Management Agent process. For example: if at time t1 (t1 < t2) the usage was a% and at time t2 it was b%, the growth % would be ((b-1)/a)%.
Metric Summary
The following table shows how often the metric's value is collected and compared against the default thresholds. The 'Consecutive Number of Occurrences Preceding Notification' column indicates the consecutive number of times the comparison against thresholds should hold TRUE before an alert is generated.
Table 1-5 Metric Summary Table
| Target Version | Evaluation and Collection Frequency | Upload Frequency | Operator | Default Warning Threshold | Default Critical Threshold | Consecutive Number of Occurrences Preceding Notification | Alert Text |
|---|---|---|---|---|---|---|---|
|
All Versions |
Every 1038 Seconds |
After Every Sample |
> |
.5 |
2 |
3 |
Agent Virtual Memory Growth is %value%%% |
Data Source
The source of the raw information is the UNIX ps command. From this, we calculate an average over four interval periods and use this as our comparison percentage (that is, how much has the virtual memory usage grown as a percentage of this average.
The Response metric reports on the availability of the Management Agent.
This metric has a value of 1 if the Management Agent is up and running.
Metric Summary
The following table shows how often the metric's value is collected and compared against the default thresholds. The 'Consecutive Number of Occurrences Preceding Notification' column indicates the consecutive number of times the comparison against thresholds should hold TRUE before an alert is generated.
Table 1-6 Metric Summary Table
| Target Version | Evaluation and Collection Frequency | Upload Frequency | Operator | Default Warning Threshold | Default Critical Threshold | Consecutive Number of Occurrences Preceding Notification | Alert Text |
|---|---|---|---|---|---|---|---|
|
All Versions |
On startup |
After Every Sample |
= |
Not Defined |
0 |
1 |
Not Defined |
Data Source
Not applicable.
User Action
If the value of this metric is not 1, the Management Agent is down and contact with the Management Agent will not exist. In such situations, the Management Agent may need to be restarted.
This category of metrics provides information on the targets that do not upload data.
The Upload Statistics metrics present information on the state of the upload manager and its performance.
This metric provides a count of the targets that are not uploading data.
Data Source
The mgmt_targets, mgmt_current_availability tables in the Management Repository.
User Action
Verify the connection between the agent and OMS to which the agent is uploading is working properly. Check for frequent agent restarts, sufficient disk space for the agent upload directory, any severe agent problems logged in agent error logs, severe problems logged in the OMS error log, loader errors logged in the System Errors page.
This metric shows the number of XML files that are in the $ORACLE_HOME/sysman/emd/upload directory waiting to be uploaded to the repository.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 62 Minutes |
Data Source
The source of this information is the Management Agent itself.
User Action
A large number of files in this directory probably indicates that there is a problem uploading files to the repository. Check the emd.trc file for upload errors and act appropriately. The cause may also be a bad network or problems on the repository end.
The Size of Files to Upload metric presents the sum of the sizes of all XML files in the upload directory of the Management Agent.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 62 Minutes |
Data Source
The source of this information is the Management Agent itself.
User Action
If this metric is large, check the upload directory. If this directory has very few files, it may be they are large. If it has many files, there may be a problem uploading data to the repository. This may be due to a bad network, bad repository or Management Agent. Check the emd.trc file in the log directory for upload error messages.
The upload rate is the average rate in KB/sec at which data is uploaded to the repository.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 62 Minutes |
Data Source
The source of this information is the Management Agent itself.
User Action
If the rate is zero or close to zero, there may be problems uploading data or collecting data (because if collections stop for some reason, we have nothing to upload). Check the log files for collection and upload messages.
These metrics provide information about the user running the Management Agent.
The name of the group the Management Agent is running under.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 168 Hours |
Data Source
The source of this metric is the UNIX id command.
The Location metric shows the directory home of the Management Agent.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 168 Hours |
Data Source
The source of this information is the ORACLE_HOME environment variable.
This metric lists the other groups the Management Agent user belongs to.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 168 Hours |
Data Source
The source of this metric is the UNIX "id" command.
The User Name metric provides information on the user that started the Management Agent process.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 168 Hours |
Data Source
The source of this data is the UNIX id command.
The metrics in the User Limit Info category provide information about the system resources available to the Management Agent.
The CoreDump metric shows the maximum size of a core dump file in 512 Kbytes blocks. A value of unlimited means that the only limit is the file system limit.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 168 Hours |
Data Source
The source of this information is the UNIX ulimit command.
User Action
This metric shows the maximum size (in 512 Kbyte blocks) of a core dump file. To decrease or increase this limit, use the UNIX ulimit system command.
This metric shows the maximum size of the Management Agent's heap in Kbytes.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 168 Hours |
Data Source
The source for this information is the UNIX ulimit system command.
User Action
This metric shows the maximum heap size (in kbytes) made available to the Management Agent. To decrease or increase this limit, use the UNIX ulimit system command.
The File metric lets you know the size of the largest single file allowed by the system the Management Agent is running on. The unit is 512 Kbyte blocks. A value of "unlimited" means that the limit is the file system limit.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 168 Hours |
Data Source
The source for this information is the UNIX ulimit system command.
User Action
This metric shows the maximum file size (in blocks) allowed by the system that the Management Agent is running on. To decrease or increase this limit, use the UNIX ulimit system command.
The NoFiles metric shows the maximum number of file descriptors that the process can have.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 168 Hours |
Data Source
The source of this information is the UNIX system call ulimit.
User Action
If this limit is small (compared to the operating system maximum), it can be changed for the Management Agent process.
This metric displays the maximum size of the Management Agent's stack in Kbytes.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 168 Hours |
Data Source
The source for this information is the UNIX ulimit command.
User Action
This metric shows the maximum size (in kbytes) of the Management Agent's stack. To decrease or increase this limit, use the UNIX ulimit system command.
The time metric represents, in seconds, the maximum CPU seconds made available to the Management Agent process by the system it is running on. A value of "unlimited" means that the CPU time available to the Management Agent is unrestricted.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 168 Hours |
Data Source
This information is obtained using the UNIX ulimit system command.
User Action
This metric shows the maximum CPU time (in seconds) made available to the Management Agent. To decrease or increase this limit, use the UNIX ulimit system command.
The Virtual Mem metric shows the maximum virtual memory size that can be occupied by the Management Agent process. If this value is "unlimited" then the only limit is the operating system limit.
Metric Summary
The following table shows how often the metric's value is collected.
| Target Version | Collection Frequency |
|---|---|
| All Versions | Every 168 Hours |
Data Source
The source of this information is the UNIX ulimit system command.
User Action
If the value of the Virtual Mem metric is too high or too low, you can change the restriction on virtual memory use by the Management Agent process using the limit UNIX command.
