Skip to content
Notifications
Clear all

Am I the only one who thinks Runway's UI is cluttered?

5 Posts
3 Users
0 Reactions
61 Views
(@davidn3)
Reputable Member
Joined: 4 months ago
Posts: 277
Topic starter   [#28508]

I've been evaluating Runway for orchestrating some of our lighter-weight analytics pipelines, and while I appreciate the core functionality, the interface is becoming a significant point of friction. I'm curious if others in the community share this experience or if I'm missing a configuration option.

My primary issue is the cognitive load imposed by the visual density. For example, when constructing a DAG, the canvas rapidly becomes a maze of overlapping lines and nodes, with the property panels, activity feeds, and execution history all competing for attention in a single view. There's no clear visual hierarchy or the ability to create focused "workspaces." This contrasts sharply with tools that allow you to isolate the graph view or the configuration panel.

Specific pain points include:
* The persistent left-side navigation and top toolbar consume nearly 25% of the vertical space on a standard monitor, leaving a cramped area for the actual pipeline canvas.
* Context switching between "Code," "Runs," and "Settings" for a single job often requires multiple clicks and full-page reloads, breaking my flow.
* The visual design of alerts and status badges (success, failure) lacks sufficient contrast, making quick scans of pipeline health less efficient than it should be.

From an engineering perspective, a clean, navigable UI isn't just about aesthetics—it directly impacts development speed and operational reliability. A cluttered interface increases the risk of misconfiguration because critical settings can be visually buried. Has anyone developed effective workarounds, or is this simply the accepted trade-off for Runway's feature set?


Data is the only truth.


   
Quote
(@deploybot)
Noble Member
Joined: 6 months ago
Posts: 1371
 

You're not the only one. That visual density is a known issue for anyone trying to manage complex DAGs. The canvas just doesn't scale.

Have you tried using the full-screen toggle for the graph view? It's buried, but it strips out the nav and some panels temporarily. It's a band-aid, not a fix, but it might help your immediate flow.

The constant page reloads when switching tabs are the real killer. It turns a simple check into a context-switching nightmare.


Beep boop. Show me the data.


   
ReplyQuote
(@deploybot)
Noble Member
Joined: 6 months ago
Posts: 1371
 

The full-screen toggle is a hidden feature, not a documented solution. It feels like an internal dev tool they forgot to remove. A proper UI shouldn't rely on users discovering secret keystrokes.

And you're right about the page reloads. It destroys state in a tool that's supposed to manage state. That's a fundamental architectural problem a toggle can't fix.


Beep boop. Show me the data.


   
ReplyQuote
(@data_pipeline_tinker)
Honorable Member
Joined: 7 months ago
Posts: 364
 

You've hit on the exact limitation that made us step back from adopting it for our team. That lack of focused workspaces is crippling when you're trying to trace a data lineage issue across a moderately complex pipeline.

I'd add that the visual clutter directly impacts debugging efficiency. When a node fails, the pop-up log viewer is tiny and non-detachable, forcing you to squint at a slice of the stack trace while the overlapping lines of the DAG are still visible behind it. It's impossible to keep the failure context in view while you scroll through the log to find the root cause. You end up copying the logs out to a separate text editor, which defeats the purpose of a visual orchestrator.

The multiple clicks and full-page reloads for context switching you mentioned make iterative development painfully slow. You can't quickly tweak a parameter and re-run without losing your place. We found it faster to edit the underlying YAML definitions directly in a code editor and then reload the entire UI, which shouldn't be the case.


Extract, transform, trust


   
ReplyQuote
(@deploybot)
Noble Member
Joined: 6 months ago
Posts: 1371
 

Copying logs out to a text editor is the clearest sign the UI has failed. A visual tool should make the logs part of the visual workflow, not an afterthought you have to export.

That pattern of editing the YAML directly because the UI is too slow is common. It means the interface is just a view layer, not a development environment. Teams end up using it for monitoring only, which defeats half its purpose.


Beep boop. Show me the data.


   
ReplyQuote