1. From the first execute object the order of execution is one by one from top to bottom (our example in the end of the first line table target is used in the second line as a single line runs our example, the table target at the end of the first line is used in the second line as a source)
or
2. N8n runs all 6 lines in parallel, so if it has a free resource, it will start from top and run lines in paralle l?
1. From the first execute object the order of execution is one by one from top to bottom (our example in the end of the first line table target is used in the second line as a single line runs our example, the table target at the end of the first line is used in the second line as a source)
or
2. N8n runs all 6 lines in parallel, so if it has a free resource, it will start from top and run lines in paralle l?
depends on whether ur lines actually share data. from what u described — table target at the end of line 1 being used as source for line 2 — those have a real dependency, so n8n runs em sequentially, top one first. the engine cant start line 2 until line 1 finishes since line 2 needs that output.
if the 6 lines were truly independent (no shared data), n8n’s still not running em in parallel — even with free resources, the execution engine processes nodes sequentially within a single workflow run. it CAN interleave (pick whichever node is ready next across branches), but theres no actual parallelism in a single execution. for real parallel ur looking at queue mode + multiple workers, or splitting into 6 separate workflows that get triggered together.
oh and ur image didnt upload — the link is pointing at ur local filesystem (file:///C:/Users/…) instead of the discourse cdn. try drag-dropping the screenshot into the post directly if u want it visible.
n8n does not run every visible branch in parallel just because they are on the canvas.
Execution follows the connected graph. A node runs when its input data is available. If one line depends on the result of another line, connect them so that dependency is explicit.
So if line 2 needs the table created or updated by line 1, make line 2 start after line 1. Do not rely on visual position.
With Execution Logic v1, n8n runs branches sequentially - your option 1 is correct. When a node connects to multiple branches, n8n completes each branch fully before starting the next one, in top-to-bottom order as they appear on the canvas.
So in your workflow, all 6 lines run one at a time: the first branch completes, then the second starts, and so on. This means it is safe to use the output of the first branch as a source in the second branch, as the first will always be finished by the time the second begins.
This is one of the key differences between v1 and v2 execution logic - v2 introduced a more efficient execution order that can change the sequence in which branches are processed, so v1 is the right choice when you have dependencies between branches like this .
