Active Jobs

Overview

The Active Jobs overview monitors the scheduled and running jobs in Webdesk. For each job and trigger, it shows whether the job is scheduled, currently running, long-running, inactive, or affected by a problem.

In a distributed installation, the information from all known scheduler nodes is merged into a shared overview. This makes it visible on which cluster node a job is executed and whether a node or its scheduler is unreachable.

A job that was triggered once via the Start job function can also appear temporarily in the overview. After execution completes, the entry disappears again, provided there is no active trigger for it.

Summary and currency

Above the table, important key figures are displayed:

• Total jobs counts the displayed job and trigger entries. A job with multiple triggers can therefore appear more than once.
• Running jobs shows the jobs currently being executed.
• Failed last executions shows how many jobs' last execution failed or was aborted due to a server restart. The value can be clicked to filter the list.
• Long-running shows the running jobs whose runtime exceeds the configured threshold. This value can also be used directly as a filter.
• Busy scheduler nodes shows nodes on which all available scheduler threads are occupied.
• Responding nodes shows the number of reachable nodes relative to the configured cluster nodes. The value opens the detailed cluster status view.
• Stale and Missing indicate nodes for which no current, or no, status at all is available.
• Cluster mode shows whether the installation is operated in a distributed manner or as a single server.
• Updated at indicates the point in time to which the displayed data refers.

With Refresh, the overview can be reloaded immediately. In addition, the page is automatically refreshed roughly once per minute. While a refresh or another action is running, no parallel refresh is started.

Scheduler and cluster status

Below the key figures, the known scheduler nodes are displayed with their utilization. The indication, for example 2/8, means that two of eight scheduler threads are currently in use. A color coding indicates the state of the node.

• A missing node is unreachable. For active jobs assigned to this node, the status Missing node is displayed in the table.
• For a stale node, the last status message is too old. The associated entries receive the status Stale node.
• For busy nodes, scheduled jobs may wait for free scheduler capacity. The status Threads busy is used for this.

The detailed cluster status view can be opened via the Responding nodes counter. It helps distinguish between a node that is genuinely missing, a delayed status message, and normal scheduler utilization.

Job table and filters

Depending on the installation and job configuration, the table contains the following information and controls:

The column filters for job name, client, trigger, cluster node, status, errors and execution state can be combined with sorting and page navigation. The page size can be adjusted in the table toolbar.

Threshold for long-running jobs

The threshold for long-running jobs (minutes) setting defines the runtime from which a running job is given the status Running long. The default value is 30 minutes. After a change, the running time is re-evaluated and the key figure and the list are updated accordingly.

The threshold is a monitoring aid and does not automatically terminate the job. For targeted control, the running jobs can first be filtered by Running long and then checked via status, runtime and job log.

Starting a job once and checking logs

To run a job once, open the job editor via the pencil icon and select Start job there. This function is available when the job engine is running. After starting, a link to the associated log is offered.

The log icons in the Status column and the icon in the Logs column lead to the execution details. This allows not only the final status but also the runtime, node and logged errors of the respective execution to be checked.

Cancelling a running job

A running job can be requested for cancellation via the cancel icon in its row. A confirmation is displayed before submission. This function is only available for jobs that support interruption.

After confirmation, a cancellation request is sent to the job service and the overview is reloaded. The message Cancellation request sent confirms the submission, not necessarily the already completed cancellation. Check the new status and the job log after the refresh.

Usage notes

• Use Updated at to check, before diagnosing, how current the displayed data is.
• When troubleshooting, start with the Last execution failed filter and then open the status icons or the job logs.
• A job can receive the status Error in last execution or Finished with errors despite completing successfully, if errors were logged during execution.
• For cluster problems, check the node display first and then the detailed cluster status view. A missing or stale node can explain why a scheduled job does not start.
• The status of the overview is a snapshot. For a complete analysis of an execution, the linked job logs are authoritative.

Felder