Work states

Work goes through a series of State changes over its lifetime.

One-time work states

For a one-time work request, your work begins in an ENQUEUED state.

In the ENQUEUED state, your work is eligible to run as soon as its Constraints and initial delay timing requirements are met. From there it moves to a RUNNING state and then depending on the outcome of the work it may move to SUCCEEDED, FAILED, or possibly back to ENQUEUED if the result is retry. At any point in the process, work can be cancelled, at which point it will move to the CANCELLED state.

Figure 1 illustrates the life of one-time work, with the events that may take it to another state.

Figure 1. State diagram for one-time work.

SUCCEEDED, FAILED and CANCELLED all represent a terminal state for this work. If your work is in any of these states, WorkInfo.State.isFinished() returns true.

Periodic work states

Success and failed states apply only to one-time and chained work. For periodic work, there is only one terminal state, CANCELLED. This is because periodic work never ends. After each run, it’s rescheduled, regardless of the result. Figure 2 depicts the condensed state diagram for periodic work.

Figure 2. State diagram for periodic work.

Blocked state

There is one final state we haven’t mentioned yet, and that is BLOCKED. This state applies to work that is orchestrated in a series, or chain of work. Work chains, and their state diagram, are covered in Chaining work.

Next Steps

In Managing work, you’ll learn more about how to manage and monitor the progress of your work.