Edit-task

File:Admin edit task.png

Editing a task

A Task is the unit of execution. It requires an existing Group and an existing Connection.

Creating a task

1. Open Author → Tasks.
2. Click Add task in the task hierarchy header.
3. Select the Group — the list shows entries as scenario - group.
4. Select the Connection. This determines which further fields appear.
5. Enter the command in the connection-specific section.
6. Click Save.

Identification

Field	Notes
ID	Assigned automatically; read-only.
Name	Mandatory. Must be unique within the group, so the same name may be reused in another group.
Enabled	When cleared, the task is not included when a processing is planned.
Group	The owning group. Changing it also changes the scenario.
Connection	How the task is executed.
Description	Free text.
Milestone	Marks the task as a milestone tracked on the Dashboard. When set, a Deadline field appears for the time it is expected to complete.

The command

The fields for entering the command depend on the connection's provider — database, script, REST, ODI, Azure, AWS or GCP each have their own section. See Connection and its sub-pages.

Three helpers sit above the command area:

• Add variable — lists the available placeholders with their meanings, so they do not have to be remembered.
• Select template — offered for OS Command tasks; inserts a predefined command skeleton.
• Error handler — offered for database tasks; wraps the statement in an exception-handling block.

Use Add variable rather than typing placeholders by hand — the list is authoritative and the names are easy to get subtly wrong.

Prioritisation and load

Field	Range	Default	Meaning
Priority	0–9999	20	Lower runs earlier.
Weight	0–9999	10	How much load the task represents, against group and connection load limits.
Max repeat	0–9999	1	Further attempts after a failure.
Wait to retry	1–9999 min	10	Interval between attempts.
Inherits priority		on	Take the priority of a dependent task where it is higher.

Weights are assigned by hand and are not adjusted automatically. If a task's resource usage changes significantly, the weight must be updated. See Resource management and Prioritization.

Scheduling

Field	Notes
Start date / End date	Validity, evaluated against the processing date.
Schedule	Daily, Weekly, Monthly, Quarterly or Yearly.
Weekdays	For Weekly — the days on which the task runs.
Day	For Monthly, Quarterly and Yearly — the day within the period. The reversed checkbox counts from the end instead of the start, which is how "last day of the month" is expressed.
Not before / Not after	A time window within the day during which the task may run.

Field	Options	Meaning
Execution mode	Workflow, Schedule	How the task participates in the flow.
Multiday mode	Group setting, Parallel, Serial, Consumed, Fully processed	Behaviour when several dates are active at once. See Parallelism.
Dependency mode	All dependencies, First dependency	Whether to wait for every dependency or continue after the first.

Warnings

• Warning after repeats — report a warning once the retry count has been exhausted.
• Warning return value — for script tasks, a return value between 0 and 255 to be treated as a warning rather than an error.

A task in Warning status has completed and does not block its dependants; see Handling failed tasks.

Dependencies

Add a dependency by typing into the Add... box and selecting from the suggestions. Entries show the task ID, group and name. Only the first 50 matches are listed, so narrow the text if the task you want does not appear.

Each entry in the list has a type button that toggles between:

• On success — the default; this task waits for the other to finish successfully
• Exclusive — the two tasks may not run at the same time

Use Exclusive for tasks that contend for the same resource — two loads into the same target table, for example — where running them together causes lock waits or deadlocks.

Inter-scenario dependencies

Ticking Inter-scenario extends the search to tasks in other scenarios. Cross-scenario dependencies are marked with a distinct icon both in the suggestion list and in the dependency list.

Use them sparingly. A dependency across a scenario boundary couples two planning events that were intended to be independent, and makes it harder to reason about why a processing is waiting.

Notifications and tags

The Notifications list receives notifications when this task fails, in addition to those assigned at group and scenario level. Notification is sent only after the last automatic attempt has failed. See Operator.

Tags are added from the Add tag... box and shown as removable chips. They have no effect on execution; see Tag.

Changes to dependencies, notifications and tags apply immediately and are not part of Save.

Practical notes

• Check the executed query after the first run. The Query tab in the Monitor console shows the statement after placeholder substitution — the quickest way to confirm the definition is right.
• Set the weight deliberately for heavy tasks. Leaving every task at the default makes the load limit meaningless as a control.
• Run "Check cyclic dependencies" after adding dependencies. A cycle introduced after a processing has been planned is not detected automatically.
• Recalculate the reference run after a change that materially alters runtime, or estimates will continue to reflect the old profile.

See also

• Task
• Dependencies
• Placeholders
• edit-group