Dependencies

cross-scheduler dependencies,dependency between schedulers,scheduler dependencies

To determine a jobstream flow and be able to forecast and predict jobstream completion statuses, AAI identifies and considers dependencies between jobs within a process flow. Whereas workflow automation engines typically can model and follow dependencies between jobs in the workflows in their scheduler instance, AAI can model job dependencies between scheduler instances of the same type and even schedulers of different types. All these can be included in one start-to-finish jobstream that models the entire workflow. With everything in one jobstream, AAI can make forecasts and predictions and track the jobstream progress for true monitoring of the entire process against the SLA deadline. The increased visibility supports triage efforts. When the entire workflow, even over more than one scheduler, is modeled in a jobstream, AAI can store and aggregate statistics that are meaningful for both monitoring, triage, and analysis.

This page includes the following:

Cross-Instance Jobstream Dependencies (AutoSys/CA7)

dependency between AutoSys and CA7

Both AutoSys and CA7 allow dependencies to be created across multiple instances. AAI also supports cross-instance dependencies in the building and viewing of the jobstreams. If an upstream dependency references a job (or jobs) that exist in a different instance and that instance is defined to AAI, the predecessor job and all of its definitional properties and event information will be available in AAI. The information is viewable in real-time monitoring, Critical path historical analysis Gantt charts, as well as in Reporting.

When viewing jobstreams on the Monitoring or Analysis Gantt charts, the line for the predecessor jobs is green when the jobstream has cross-instance dependencies.

On Monitoring and Analysis screens, jobstreams with cross-instance dependencies look basically the same as those without, except that the line that indicates the predecessor job(s) will be green, rather than the standard blue.

You may also choose to turn those lines on separately (by removing the critical path lines and turning on cross instance dependency lines) to more easily isolate and distinguish cross-instance dependencies in a jobstream.

Important

For AAI administrators: To enable cross-scheduler dependency calculations, predictions, visualizations, and simulations for AutoSys jobstreams, you must select the Next Generation option (the default) as the Prediction Model when adding the AutoSys scheduler. You must do this when originally adding the scheduler because you cannot modify this setting afterward.

Cross-Instance(s) when AutoSys/CA7 Instance Not Defined

Cross-instance dependencies will not be continued across the instances if the AutoSys/CA7 instance where those jobs are defined is not connected to AAI. Instead, you will simply see an indication that there is a cross instance job upstream and there will be no run information for that job. Visually, this will be represented in AAI by the No Run icon, which is a rectangle box with an intersecting slash. A green line will point from the No Run icon to the upstream job.

To help you discover cross-instance dependencies in your environment where the referenced instance is not yet defined to AAI, refer to the Conversion Report that is generated when defining a new instance. You may locate the Conversion Report on the Admin tab by clicking on a scheduler, and then the View Log button. In the figure above, you will see that a scheduler (danang) is added to AAI, generating a Conversion Report that lists jobs contained in danang, which references a job in another instance (ACE), currently not defined to AAI. You will have to add that instance (ACE) for the upstream job(s) referenced to appear.

Note:

If you add an instance that is referenced by a job in another instance, and the one that you are adding has the same instance name as another, AAI will require you to resolve that reference in the Edit Scheduler dialog (x-instances tab).

Job Dependencies in Universal Schedulers

dependency between schedulers,dependency between universal schedulers

In addition to end-to-start dependencies in jobstreams, universal schedulers also support start-to-start job dependencies. An end-to-start dependency means that a job starts when its predecessor ends. A start-to-start job dependency means that a job starts when its predecessor starts.

To understand how AAI calculates the latency in either type of dependency, you must be acquainted with the following concepts:

  • Finish

    Time when a dependency condition is met. For end-to-start dependencies this is when the predecessor job ends. For start-to-start dependencies this is when the predecessor job starts.

  • Finish to start latency

    Lapse of time between Finish and the successor start

    • For end-to-start dependencies, the Finish to start latency is calculated as follows:

      Successor start time - Predecessor end time

    • For start-to-start dependencies the Finish to start latency is calculated as follows:

      Successor start time - Predecessor start time

    Note:

    (ESP Workload Automation only) ESP jobs always have a parent application. For calculating their Finish to start latency, the minimum latency is used. One of the following calculation methods is used:

    • min(successor start time - parent start time)

    • min(successor start time - Finish)

Graphic Representation

You can easily recognize start-to-start and end-to-start dependencies in the Gantt chart.

The arrow for a start-to-start dependency goes from the beginning of the successor job to the beginning of the predecessor job. The arrow for an end-to-start dependency goes from the end of the successor job to the beginning of the predecessor job.

Screenshot showing part of a jobstream where both types of dependencies are displayed

Predictions and Forecast Calculations

For both types of dependencies, the average Finish to start latency is always considered in forecast calculations.

Universal to Universal Cross Instance Dependencies

AAI depicts the dependencies between jobs coming from different universal schedulers in the same way as within one scheduler. For example, a job in Universal Scheduler A drops a file on a specific directory that triggers a job in Universal Scheduler B to run. AAI displays this dependency, traces the critical path and connects everything for purposes of analytics. If any of the jobs in the Universal Scheduler A fails or has a delay, AAI will send a notification to the user indicating that there's a risk of missing the SLA on Universal Scheduler B.

Important

The only prerequisite for AAI to be able to depict universal to universal cross instance dependencies is that custom conditions are available that represent those dependencies. For more information, see

Example

The screenshot below shows a very simple custom dependency between two universal schedulers. A green arrow drawn between the predecessor and the successor depicts the universal to universal dependency.

Click to expand:

On the View Settings dialog you define whether the green arrow should be displayed or not. To display it, select the Cross-instance Dependencies checkbox.

Supported Cross-Scheduler Dependencies

AAI supports cross-instance dependencies between the following schedulers:

  • Control-M

  • Control-M (z)

  • ESP

  • Automic Automation

See also: