Runtime Management

In the previous step, you looked at workflow templates, which allow you to quickly create a workflow by starting from a pre-built state.

When it comes to workflows, the great place to get an overall picture of your organisation health is Power Platform Admin Center.

You can filter the dashboard by instance at a specific timeframe. Once you get an overall picture across the dashboard, you can then use various tools within an individual instance to monitor and troubleshoot the workflow.

For background workflows, you can access information about the system jobs in several places within the application.

You can navigate to system jobs on the settings and this list will include all types of system jobs. You will need to filter the records to those where the system job type is workflow. Then you can navigate directly from the workflow process definition, open the background workflow definition and go to the process session tab.

Note that the background workflows generates system job records to track their status. If the workflow is configured to delete successful jobs, then the log entry will be deleted shortly after the execution completes and the list will only contain the jobs that are still running, have been postponed or cancelled or failed.

When the job generates an error, it is placed in a suspended state with the reason “waiting”.

Depending on the current status of the job, the following actions are available:

• Cancel will stop the job execution with the cancel status reason.
• Postpone will place the job in waiting state until the specified date and time.
• Pause will suspend the job indefinitely and manual resume will be required.
• Resume can be applied when the job is in waiting state; that is, the job was postponed, paused, or one of the steps has failed.

Real-time workflows and actions do not use system job records because they run immediately. Any errors will be displayed to the user in the application with the heading business process error. And there is no log for successful operations

Troubleshooting

If a workflow or workflows are failing, a good place to start is Power Platform Admin Center analytics, just to get an overall picture.

Here you can analyse system job logs and individual records to see what’s causing the problem and look for the jobs failing on a regular basis. If a synchronous system job fails several times consecutively, Common Data Service starts to postpone the execution of the job for longer and a longer period of time to allow administrators to investigate and resolve the issue.

Once the job starts succeeding again, it will resume executing normally.

You can open individual job records to see other details. Often, errors come from plug-ins and custom workflow extensions. You can pass that information and error logs to the developers.

You can also detect recursive jobs by looking for multiple consecutive sessions and then a failer with a specific error message referring to an infinite loop.

Finally, don’t just resume failed workflows without any remedial action. Resume the job that is in a waiting state because an error will usually result in the same error unless some steps are taken to resolve the root cause, for example, missing critical information.