Hi,
I run n8n self-hosted n8n and the executions screen has been a pain point for me in three ways:
There’s no way to sort by run time. When something is slow I want to find the longest executions, but the list only shows newest first, so I end up scrolling and eyeballing the run time column.
The workflow filter is single-select. I have groups of related workflows and I’d like to see their executions together without clicking through them one at a time.
The “highlighted data” filter only takes one key/value pair. I save things like customer id and order type on executions, and I can’t filter by both at once. While digging into this I noticed something odd: the API actually accepts an array of pairs, but the query builder only applies the filter when there’s exactly one pair send two and you silently get completely unfiltered results back. That looks like a bug rather than a design decision.
I’ve already built these three things on a fork to solve my own problem: sort by runtime , a multi-select workflow filter, and support for multiple highlighted-data pairs with AND logic (which also fixes the silent no-op above). I used AI tooling for a good part of the implementation, per your contributing guidelines I’m mentioning that upfront.
Before I open anything: would you accept this? And if yes, would you prefer one PR or three separate ones
Happy to adjust the approach based on your feedback.
Thanks