New in xyOps v1.1.9: Run Workflow Again with advanced replay controls

[jhuckaby]

New in xyOps v1.1.9: Run Workflow Again with advanced replay controls

Ever had a workflow finish an expensive data-gathering step, only to hit a temporary failure further down the line? Running everything again can mean repeating work that already succeeded.

I'm happy to share an upgrade to the Run Again button in xyOps v1.1.9: three replay modes for completed workflows. You can run all jobs fresh, automatically freeze successful nodes, or choose exactly which nodes should reuse their previous results.

This builds on Replay Previous Jobs, bringing replay controls directly into the Run Workflow Again dialog.

What does "freeze" mean?

A frozen node reuses its previous completed job result when the workflow reaches it. Its saved output data, files, tags, and completion status feed into the workflow, without executing that job again. Unfrozen nodes run fresh jobs when reached.

Each rerun starts a new workflow job using the previous job's workflow map. Your freeze choices apply to that new run, leaving the saved workflow event and original job history intact.

Open a completed workflow job, click Run Again, and choose a Replay Mode:

Run All Jobs

The default option gives you a fresh execution of every Event and Job node the workflow reaches. It clears any previous replay settings on those nodes, including ones carried over from an earlier rerun.

Choose this when you want to fetch current data, repeat the full process, or verify that everything works together. Normal workflow conditions still apply, so a branch only runs if its connections lead there.

Freeze Successful Nodes

This mode automatically reuses results from successful nodes. Nodes that failed, or never ran, launch fresh jobs when the workflow reaches them.

For example, imagine a workflow with three steps: Generate Data, Process Data, and Final Report, connected with On Success wires. Generate Data succeeds, Process Data fails, and Final Report is never reached.

Select Freeze Successful Nodes and run it again. Generate Data supplies its saved output immediately, Process Data gets another attempt, and, if it succeeds, Final Report can run normally.

That's useful after a temporary API outage, an unavailable service, or another problem you've resolved outside the workflow. You can try the failed work again while keeping the successful results you already have.

Frozen nodes show a Replay badge in the graph. Here, Generate Data is frozen while the other two nodes remain available to run fresh:

Freeze Custom Nodes...

For more control, choose Freeze Custom Nodes... and pick from the Nodes to Freeze menu. It lists Event and Job nodes with completed jobs. Selected nodes reuse their previous results; unselected nodes run fresh jobs when reached.

This is especially handy when you want to:

• Repeat downstream processing with the same input. Freeze a database query or file-generation step, then rerun the processing or reporting nodes that consume its output.
• Avoid repeating expensive requests. Reuse a saved AI response or rate-limited API result while exercising the rest of the workflow.
• Reproduce an error path. Freeze a node with a known failure result to exercise downstream error-handling nodes without recreating the original failure.

Custom freezing preserves failures too. If you want a failed node to make a new attempt, leave it unfrozen.

A few useful details

For nodes invoked multiple times by Split, Multiplex, or Repeat controllers, freezing replays the node's previous jobs in sequence. Freeze Successful Nodes only freezes a node if all of its relevant previous jobs succeeded. If any failed, the whole node runs fresh. Attempts superseded by a retry are excluded from selection.

Those saved results are replayed in order, without matching them to current split items or servers. If your inputs or targets have changed, leave the affected nodes unfrozen so they process the current data. The multi-job replay documentation covers this in more detail.

Actions attached directly to a frozen node, such as emails or web hooks, are not executed again. Downstream connections still respond to the saved result, including success, error, and tag conditions. Previous jobs also need to remain stored and accessible to your account.

To try it, update to xyOps v1.1.9 or later, open a completed workflow, click Run Again, select your mode, and click Run Now. See Run a Workflow Again for the full details.

I'd love to hear if this feature helps you!