Skip to main content

Support Operations on Unmeshed

This page covers common support actions you can take as a user of the Unmeshed platform around process executions. These operations are useful for monitoring, debugging, and managing workflows.

1. Resume Execution​

If a process ends in a FAILED or other terminal states without completing, you can choose to Resume it. When resumed:

  • The last failed steps are retried and any pending to schedule steps are rescheduled.
  • The workflow continues forward until completion if no further errors are encountered

This is particularly helpful for handling transient errors like network issues, temporary system outages, or flaky third-party APIs.

Resume Execution

tip

Tip: You can configure Error Policies on steps to auto-retry failures multiple times, enabling self-healing execution flows without manual intervention.

2. Resume with Latest Definition​

This works just like Resume, but instead of using the original process definition, it uses the latest available version.

  • Useful when you've made a small fix or correction to a step definition.
  • Can be used to resume from the point of failure using updated logic.
warning

⚠️ Caution: Only use this if the structure of the updated process definition is compatible with the current execution. Structural changes (e.g., new/removed steps) may break the resumed process.

Resume with Latest Definition

3. Rerun with Current Definition​

This option allows you to rerun the process from the beginning with:

  • The same definition used during the original execution
  • The same input payload

You can use this operation on any process regardless of its state — FAILED, COMPLETED, TERMINATED, or even RUNNING (useful for troubleshooting).

This is also useful when you want to just re-do the same set of operations for any reason.

Rerun with Current Definition

tip

You can bulk re-run previous executions to replay transactions to leverage an updated API endpoint or underlying worker function. It's very useful in long term support of functions where you might have to replay 1000s of transactions after a fix.

4. Rerun with Latest Definition​

Similar to the above, this will rerun the process with the latest process definition, but still using the same inputs as the original execution.

This is useful for rerunning operations after you've made improvements or fixes to the workflow logic.

Rerun with Latest Definition

tip

Bulk run using API is supported and can handle 1000s of transactions concurrently for replaying.

5. Terminate Running Process​

If you notice a process is stuck in running state and if you want to terminate it, you can use this option. This will stop the execution of the process and cancel any remaining scheduled steps.

This is useful for stopping a broken flow or to stop something from progressing if you know it won't work.

Terminating Running Process

tip

Bulk run using API is supported and can handle 1000s of transactions concurrently for termination.

6. Marking as Reviewed​

note

Did you know that you can track errors with Unmeshed like a ticketing tool? Mark as reviewed comes in handy to mark a previously FAILED or TERMINATED process as something that you have already debugged. This is useful to avoid looking at the same transaction over and over again.

Marks a process as "REVIEWED". It doesn't affect any changes to the execution except updating the state.

Marking Process as Reviewed

tip

Bulk run using API is supported and can handle 1000s of transactions concurrently for termination.