Table of Contents |
---|
Introduction
This article gives an overview of the JobScheduler Monitoring Interface. This solution provides an efficient means of of monitoring JobScheduler objects such as Jobs, Job Chains and Orders and forwarding notifications to system Monitors such as op5, Nagios or icing. This solution is available with JobScheduler General Availability Release 1.8 onwards.
The most important features of this solution are:
- JobScheduler: carries out a two step process around the interface:
- Detecting errors and other events: A Job running at regular intervals - typically every 2 minutes - analyses the History Log information recorded by the JobScheduler in the database. This job is configured not only to filter but also to analyze the log information for the Job Scheduler objects being monitored. The information noted is typically whether tasks have been successfully completed or whether errors or warnings have been logged. This job then writes this information in a separate Notifications database table.
- Sending alerts: A second Job is responsible for sending the alerts to the relevant System Monitor. This job is also run at regular intervals, analyzing the Notifications database tables. It then carries out a predefined action for each item it finds in the table. Typical actions would be informing a particular monitor that a particular type of event has occurred, such as the successful completion of an order, a job ending in error or whether error recovery is being attempted.
- JobScheduler: The solution architecture allows analysis of the Log History of more than one JobScheduler using the database specified. It may also be configured to monitor more that one database.
- System Monitors: the JobScheduler is able to connect to more than one System Monitor at the same time.
Monitoring Definitions
The following definitions apply for the monitoring systems:
Definition | Description |
---|---|
System Monitor | A System Monitor is an instrument to inform a Service Desk (e.g. 1st Level Support) about incidents in IT systems. It does not analyze incidents, but merely information about incidents, in order to be able to forward and scale this information. |
Passive Checks | Passive checks are sent remotely from an external host (from the point of view of the System Monitor) to the Monitor. Otherwise, checks that are carried out periodically by the System Monitor itself are called active checks. |
Alerting | An Alert is a message about an event. An Alert does not provide all the information about an event, but it informs about the existence of the event. An Alert can be either positive or negative. |
Notification | The notification of a specific Alert. Notification will not be provided for every Alert, just the ones that are so configured will be notified. Notifications are therefore a subset of the Alerts and can also be either positive or negative. |
Acknowledgment | Is the confirmation of an alert and it has the meaning that the alert has been seen and/or is known and that appropriate action is being taken. An acknowledgment is always manually executed. This means that there is always someone that has realized there is a Critical service and this person acknowledges the services (usually by the Service Desk or 1st Level Support). It is never an automatized step. |
Benefits
The benefits of the new solution are:
- Flexible implementation:
Changes to your existing JobScheduler configuration (Jobs, Job Chains, etc.) are not required to get this solution working. You add the Job Chains required for the monitoring but do not have to modify your current ones. - Monitor independence:
The whole architecture lies on the JobScheduler side and the solution is therefore independent of the monitor that the Alerts are sent to. The solution works for every monitor that can receive passive checks. - Workload-independent:
Processing of Jobs and Job Chains in JobScheduler is not affected or modified by the monitoring, neither from the point of view of performance nor that of stability. - Clearly defined information flow:
This solution allows the information to be made available to the System Monitors to be exactly configured. Detailed log information from monitored Job Chains can be sent as a Passive Check to the relevant Monitoring Service if required. - Error Prioritization
Errors of a critical nature are immediately recognized in the System Monitor. The JobScheduler has initially access to all the log information and can be configured to filter this information very exactly before forwarding it to the relevant System Monitor Service. This feature allows the Service Desk to be able to set priorities immediately when, for example, recovering errors: it is unlikely that a performance error would be given the same priority as an error in document processing. This feature is illustrated in the following diagram:
Functionality
Functionality | Description |
---|---|
Job Chain and Order Monitoring | This solution allows Job Chains to be monitored by way of the Orders that trigger these Job Chains. |
History Notifications | Not only can critical alerts be monitored, but also positive ones. The history of a specific service can be monitored to see exactly if a specific work-flow has been executed and what result it gave. |
Performance measurement (Timer) | Timers can be used to measure the performance of Job Chains. These can be used to send a warning alert to a System Monitor if a Job Chain takes more that a predefined time to complete. |
Acknowledgment | Acknowledgments sent in response to critical alerts sent out by a System Monitor can be used to add Orders to the JobScheduler, so that the JobScheduler does not send more notifications about a service to the System Monitor. |
Monitoring example - op5 Monitor
The following example illustrates use of the JobScheduler Monitoring interface with the op5 Monitor. In the example, 3 checks (in op5 Monitor they are called services) have been defined for the JobScheduler monitoring. Different Job Chains in JobScheduler can send notifications to the same check, so that it is not necessary to create checks for each individual Job Chain, which could become extremely complex. Instead, results have been grouped in three categories:
...
Installation
See JobScheduler Monitoring Interface - Prerequisites and Installation
Configuration
JobScheduler - SystemMonitorNotification files
Location: <scheduler_install>/config/notification
File | Description |
---|---|
SystemMonitorNotification_v1.0.xsd | The XML Schema file that defines which values are allowed in your XML files for the JobScheduler monitoring. That means, you just have to modify your |
SystemMonitorNotification_<MonitorSystem>.xml | Configuration file for each System Monitor.
|
| Configuration file for all System Monitors.
This file is optional and contains the definitions of the |
SystemMonitorNotification Elements
The configuration element descriptions are organized into the following major categories:
Element | Element description | Description |
---|---|---|
SystemMonitorNotification | Top Level Element | Configuration for notifications to a system monitor |
Notification | Once or more inside a SystemMonitorNotification element | Specifies a system monitor notification that includes a command line invocation and the JobScheduler objects |
Timer | Optional, once or more inside a SystemMonitorNotification element | Performance measurement definition |
SystemMonitorNotification
SystemMonitorNotification
support the following attributes:
...
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
<SystemMonitorNotification system_id="OP5"> ... |
SystemMonitorNotification / Notification
The following elements may be nested inside a Notification
element:
Element | Element description | Description |
---|---|---|
NotificationMonitor | Once inside a Notification element | Specifies the System Monitor interface that is being used for messages: either by a Plug-in Interface or by command line invocation |
NotificationObjects | Once inside a Notification element | Specifies the Job Chain and the Timer definitions |
SystemMonitorNotification / Notification / NotificationMonitor
NotificationMonitor
supports the following attributes:
...
Element | Element description | Description |
---|---|---|
NotificationInterface | Optional, once inside of NotificationMonitor element | Plugin Interface to be executed for System Monitor notification |
NotificationCommand | Optional, once inside of NotificationMonitor element | Command line to be executed for System Monitor notification |
SystemMonitorNotification / Notification / NotificationMonitor / NotificationInterface
NotificationInterface
support the following attributes:
...
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
... <NotificationInterface monitor_host="monitor_host" monitor_port="5667" monitor_encryption="XOR" service_host="service_host"><![CDATA[ scheduler id=%MON_N_SCHEDULER_ID%, history id=%MON_N_ORDER_HISTORY_ID%, job_chain=%MON_N_JOB_CHAIN_NAME%(%MON_N_ORDER_ID%), step =%MON_N_ORDER_STEP_STATE%, error=%MON_N_ERROR_TEXT% ]]></NotificationInterface> ... |
SystemMonitorNotification / Notification / NotificationMonitor / NotificationCommand
NotificationCommand
support the following attributes:
...
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
... <NotificationCommand><![CDATA[ echo scheduler id=%MON_N_SCHEDULER_ID%, history id=%MON_N_ORDER_HISTORY_ID%, job_chain=%MON_N_JOB_CHAIN_NAME%(%MON_N_ORDER_ID%), step =%MON_N_ORDER_STEP_STATE%, error=%MON_N_ERROR_TEXT% > D://errors.txt ]]></NotificationCommand> ... |
SystemMonitorNotification / Notification / NotificationObjects
One of the following elements must be nested inside a NotificationObjects
element:
...
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
<SystemMonitorNotification system_id="OP5"> <Notification> <NotificationMonitor service_name_on_error="JobScheduler Monitoring Errors"> ... </NotificationMonitor> <NotificationObjects> <!-- Send the job chain error, occurrent in the "test/my_jobchain" job chain, to the "JobScheduler Monitoring Errors" service. --> <JobChain name="test/my_jobchain" /> </NotificationObjects> </Notification> </SystemMonitorNotification> |
SystemMonitorNotification / Notification / NotificationObjects / JobChain
JobChain
support the following attributes:
...
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
... <JobChain notifications="2" name="test/my_jobchain"/> ... <JobChain scheduler_id="scheduler_4444" /> ... <JobChain scheduler_id="scheduler_4444" name="^(test/my)" /> ... <JobChain name="test/my_jobchain" step_from="200"/> ... <JobChain name="test/my_jobchain" step_to="500"/> ... <JobChain name="test/my_jobchain" step_from="300" step_to="300"/> ... <JobChain name="test/my_jobchain" excluded_steps="200;300"/> ... |
SystemMonitorNotification / Notification / NotificationObjects / Timer
Timer support the following attributes:
...
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
<SystemMonitorNotification system_id="OP5"> <Notification> <NotificationMonitor service_name_on_error="JobScheduler Monitoring Error"> ... </NotificationMonitor> <NotificationObjects> <!-- Send the job chain error, occurrent in the "test/my_jobchain" job chain, to the "JobScheduler Monitoring Errors" service. --> <JobChain name="test/my_jobchain" /> </NotificationObjects> </Notification> <Notification> <NotificationMonitor service_name_on_error="JobScheduler Monitoring Performance"> ... </NotificationMonitor> <NotificationObjects> <!-- Send the performance check error, occurrent in the "test/my_jobchain" job chain, to the "JobScheduler Monitoring Performance" service. Send of the performance check error to the "JobScheduler Monitoring Performance" service will be ignored when the "test/my_jobchain" has the job chain error (default notify_on_error = false). --> <Timer name="my_timer" /> </NotificationObjects> </Notification> <Timer name="my_timer"> <JobChain name="test/my_jobchain" /> </Timer> </SystemMonitorNotification> |
SystemMonitorNotification / Timer
The following elements must be nested inside a Timer
element:
...
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
... <Timer name="my_timer"> ... |
SystemMonitorNotification / Timer / JobChain
JobChain
support the following attributes:
...
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
... <JobChain scheduler_id="scheduler_4444" /> ... <JobChain scheduler_id="scheduler_4444" name="^(test/my)" /> ... <JobChain name="test/my_jobchain" step_from="200"/> ... <JobChain name="test/my_jobchain" step_to="500"/> ... <JobChain name="test/my_jobchain" step_from="300" step_to="300"/> ... |
SystemMonitorNotification / Timer / Minimum
The following elements must be nested inside a Minimum
element:
...
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
... <Timer name="my_timer"> ... <Maximum><Script language="javascript"><![CDATA[1000]]></Script></Maximum> </Timer> ... |
SystemMonitorNotification / Timer / Maximum
The following elements must be nested inside a Maximum
element:
...
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
... <Timer name="my_timer"> ... <Minimum><Script language="javascript"><![CDATA[1000]]></Script></Minimum> </Timer> ... |
SystemMonitorNotification / Timer / Minimum|Maximum / Script
Script
support the following attributes:
...
- fixed value
- calculation based on the job/order parametes
Fixed value
Fixed value is the duration time in seconds for the specific Minimum
or Maximum
definition
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
... <Script language="javascript"><![CDATA[1000]]></Script> ... |
Calculation
The calculation result is the time in seconds for the specific Minimum
or Maximum
definition.
...
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
<?xml version="1.0" encoding="ISO-8859-1"?> <job title="Sample Job with Store Result Monitor" order="yes" stop_on_error="no" tasks="1"> <params> <!-- set the scheduler_notification_result_parameters parameter --> <param name="scheduler_notification_result_parameters" value="file_size"/> </params> <!-- calculate and create the new order parameter if necessary --> <script language="javascript"><![CDATA[ function spooler_process(){ var order = spooler_task.order; var params = spooler.create_variable_set(); params.merge(spooler_task.params); params.merge(order.params); // parameter scheduler_file_path was set in the previous job chain step var file = new java.io.File(params.value("scheduler_file_path")); var fileSize = file.length()/1024; order.params.set_var("file_size",fileSize.toString()); return true; }]]> </script> <!-- set the com.sos.scheduler.notification.jobs.result.StoreResultsJobJSAdapterClass as monitor --> <monitor name="notification_monitor" ordering="1"> <script java_class="com.sos.scheduler.notification.jobs.result.StoreResultsJobJSAdapterClass" language="java"/> </monitor> <run_time /> </job> |
Message
Usage
The Message can be configured on the following parent nodes as CDATA element :
...
Example: <![CDATA[ scheduler id = %MON_N_SCHEDULER_ID% ]]>
Variables
All variables must be defined by using of the %<variable name>%
syntax.
...
- Table variables.
- Service variables.
- OS environment variables.
Table variables
Expand | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Table of the history of steps of processed orders.
|
...
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
timer name = %MON_C_NAME%, text = %MON_C_CHECK_TEXT% |
Service variables
Expand | ||||||||
---|---|---|---|---|---|---|---|---|
| ||||||||
|
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
service name = %SERVICE_NAME% |
OS environment variables
All existing system variables can be defined by message with the syntax like %<variable name>%
(Windows/Unix)
.
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
%TEMP%/test.exe |
Examples
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
scheduler id=%MON_N_SCHEDULER_ID%, history id=%MON_N_ORDER_HISTORY_ID%, job_chain=%MON_N_JOB_CHAIN_NAME%(%MON_N_ORDER_ID%), step=%MON_N_ORDER_STEP_STATE%, error=%MON_N_ERROR_TEXT% |
...
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
name = %MON_C_NAME%, scheduler id=%MON_N_SCHEDULER_ID%, history id=%MON_N_ORDER_HISTORY_ID%, job_chain=%MON_N_JOB_CHAIN_NAME%(%MON_N_ORDER_ID%), steps(%MON_C_STEP_FROM% to %MON_C_STEP_TO%), check = %MON_C_CHECK_TEXT% |
Notification environment variables
The default com.sos.scheduler.notification.plugins.notifier.SystemNotifierProcessBuilderPlugin
plugin used by the SystemMonitorNotification / Notification / NotificationCommand
element sets the following variables as environment variables:
...
These variables can be used when the NotificationCommand calls the notification client not directly, but a shell script, that make the logical implementation for sending of the notification messages.
Table variables
Expand | ||
---|---|---|
| ||
All table variables (see
e.g.:
|
Service variables
Expand | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||
|
...
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
1) configured command in the SystemMonitorNotification_<MonitorSystem>.xml file <NotificationCommand><![CDATA[C:/Temp/command.cmd]</NotificationCommand> 2) content of the C:/Temp/command.cmd file rem Note: "> C:/Temp/command_output.txt" used to simulate the starting of the notification client rem echo %SCHEDULER_MON_SERVICE_NAME%:%SCHEDULER_MON_SERVICE_STATUS%:%SCHEDULER_MON_SERVICE_MESSAGE_PREFIX% history id = %SCHEDULER_MON_TABLE_MON_N_ORDER_HISTORY_ID% > C:/Temp/command_output.txt |
Examples
Examples OP5
NotificationInterface
Here is an except of an XML file used for notifying a specific System Monitor (OP5 Monitor) and using NotificationInterface:
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
... <!-- monitor_host The hostname or ip address of System Monitor host monitor_port The TCP port that the System Monitor would listen to monitor_encryption Encryption algorithm service_host The host that executes the passive check. The name must match the corresponding setting in the System Monitor %MON_N_SCHEDULER_ID% See explanation "Table variables" ... --> <NotificationInterface monitor_host="monitor_host" monitor_port="5667" monitor_encryption="XOR" service_host="service_host"><![CDATA[ scheduler id=%MON_N_SCHEDULER_ID%, history id=%MON_N_ORDER_HISTORY_ID%, job_chain=%MON_N_JOB_CHAIN_NAME%(%MON_N_ORDER_ID%), step =%MON_N_ORDER_STEP_STATE%, error=%MON_N_ERROR_TEXT% ]]></NotificationInterface> ... |
NotificationCommand
Here is an except of an XML file used for notifying a specific System Monitor (OP5 Monitor) and using NotificationCommand on Windows:
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
... <!-- service_host The host that executes the passive check. The name must match the corresponding setting in the System Monitor. monitor_host The hostname or ip address of System Monitor host. %SERVICE_NAME% See explanation "Service variables" %SERVICE_STATUS% See explanation "Service variables" %SERVICE_MESSAGE_PREFIX% See explanation "Service variables" %MON_N_SCHEDULER_ID% See explanation "Table variables" ... NotificationCommand after substitution (error case): <![CDATA[echo service_host:JobScheduler Monitoring Errors:2:ERROR scheduler id=scheduler_4444, history id=123, job_chain=test/my_jobchain(order_id), step=100, error=error occurred | D:\nsca\send_nsca.exe -H monitor_host -c D:\nsca\send_nsca.cfg -d : ]]> NotificationCommand after substitution (recovery case): <![CDATA[echo service_host:JobScheduler Monitoring Errors:0:RECOVERED scheduler id=scheduler_4444, history id=123, job_chain=test/my_jobchain(order_id), step=100, error=error occurred | D:\nsca\send_nsca.exe -H monitor_host -c D:\nsca\send_nsca.cfg -d : ]]> NotificationCommand after substitution (success case): <![CDATA[echo service_host:JobScheduler Monitoring Success:0:scheduler id=scheduler_4444, history id=123, job_chain=test/my_jobchain(order_id), step=100, error= | D:\nsca\send_nsca.exe -H monitor_host -c D:\nsca\send_nsca.cfg -d : ]]> --> <NotificationMonitor service_name_on_error="JobScheduler Monitoring Errors" service_name_on_success="JobScheduler Monitoring Success"> <NotificationCommand><![CDATA[echo service_host:%SERVICE_NAME%:%SERVICE_STATUS%:%SERVICE_MESSAGE_PREFIX%scheduler id=%MON_N_SCHEDULER_ID%, history id=%MON_N_ORDER_HISTORY_ID%, job_chain=%MON_N_JOB_CHAIN_NAME%(%MON_N_ORDER_ID%), step=%MON_N_ORDER_STEP_STATE%, error=%MON_N_ERROR_TEXT% | D:\nsca\send_nsca.exe -H monitor_host -c D:\nsca\send_nsca.cfg -d : ]]> </NotificationCommand> </NotificationMonitor> ... |
Examples Zabbix
NotificationCommand
Here is an except of an XML file used for notifying a specific System Monitor (Zabbix Monitor) and using NotificationCommand
Code Block | ||||||
---|---|---|---|---|---|---|
| ||||||
... <!-- zabbix_sender Zabbix sender installed on the JobScheduler host localhost Hostname of the zabbix server Zabbix_server JobScheduler Agent name(host name) that registred on Zabbix samples.job1 Item key of zabbix (replace "/" to "." of JOB_NAME %MON_N_ERROR_TEXT% See explanation "Table variables" --> <NotificationCommand> <![CDATA[zabbix_sender -z localhost -s zabbix_server -k samples.job1 -o %MON_N_ERROR_TEXT%]]> </NotificationCommand> ... |
JobScheduler - Job Chains
...
Status colour Yellow title work in progress
Use Cases
Recoverable Errors
Initial Situation: A Job Chain is triggered by directory monitoring. That is, when a certain file comes in a monitored folder, the Job Chain starts.
...
- XML
CheckConfigurationHistory.xml
: Indicate the ID of the JobScheduler and the name of the Job Chain you want to monitor. - XML
SystemMonitorNotification.xml
: Specify the name of the Service (in the System Monitor) and specify that it is about aservice_name_on_error
since you want to have the control when the Job Chain ends in an error. - System Monitor: Services in the System Monitor have to be configured and named the same way as in the XML file above
SystemMonitorNotification.xml
.
Workflow Execution takes too long
Initial Situation: A Job Chain is triggered and it could not end, it hanged in a step, taking then longer than expected.
...
- XML
CheckConfigurationHistory.xml
: As in the example above, indicate the ID of the JobScheduler and the name of the Job Chain you want to monitor. Moreover, specify the timer for this specific job chain and the function to calculate the expiration time for the timer. - XML
SystemMonitorNotification.xml
: As in the example above, specify the name of the Service (in the System Monitor) and specify that it is about aservice_name_on_error
since you want to have the control when the Job Chain ends in an error. Moreover and essential for this particular case, specify how many times the timer should notify your System Monitor about the expiration of a timer. - System Monitor: As in the example above, Services in the System Monitor have to be configured and named the same way as in the XML file above
SystemMonitorNotification.xml
.
SFTP connection refused
Initial Situation: There is a Job Chain that uses SFTP for transferring files. You have a setback configured in this step of the Job Chain, so that if the connection to the SFTP server fails, this step is retried after some time.
...
- XML
CheckConfigurationHistory.xml
: As in the example above, indicate the ID of the JobScheduler and the name of the Job Chain you want to monitor. - XML
SystemMonitorNotification.xml
: As in the example above, specify the name of the Service (in the System Monitor) and specify that it is about aservice_name_on_error
since you want to have the control when the Job Chain ends in an error. Moreover and very important in this case, specify how many times this Job Chain should notify your System Monitor about the error connecting to the SFTP Server. You can usestep_from
andstep_to
for that in order to reduce the number of notifications for this specific step. - System Monitor: As in the example above, Services in the System Monitor have to be configured and named the same way as in the XML file above
SystemMonitorNotification.xml
.
Thresholds
Initial Situation: For example, a specific number of Workflow Executions have to be executed successfully till some specific time. That is, a specific value has to be monitored in order to determine if this quote was reached.
...
- XML
CheckConfigurationHistory.xml
: As in the example above, indicate the ID of the JobScheduler and the name of the Job Chain you want to monitor. - XML
SystemMonitorNotification.xml
: Specify the name of the Service (in the System Monitor) but now specify that it is about aservice_name_on_success
since you want to have the control when the Job Chain ends in an success, and not only when it ends on error. - System Monitor: As in the example above, Services in the System Monitor have to be configured and named the same way as in the XML file above
SystemMonitorNotification.xml
.
Acknowledgement
Initial Situation: An alert for a Service has been sent to the System Monitor and a Mail has been sent to the Service Desk (Support Team) notifying about it.
...