Workflows and Strategies

Practical approaches for working with complex graphs

The basics of moving around (pan, zoom, centre view) are covered in the Interface Overview. This page focuses on strategies for working efficiently once graphs get large or complex.

Taming large graphs

A graph with hundreds of nodes can feel overwhelming at first. The key is to reduce what you’re looking at before trying to understand it.

Start by isolating a section. Right-click a node of interest and isolate it with relations. This strips the graph down to that node and everything it connects to. From there, you can expand outward by adding related nodes, or further narrow with exclusion filters.

Once you’ve built a useful view, bookmark it. Bookmarks are cheap and fast to create (Ctrl+B / Cmd+B). Think of them as saved perspectives on your data. A “database layer” bookmark, a “failing services” bookmark, a “team X ownership” bookmark. Switch between them instantly instead of rebuilding filters each time.

Working with deep hierarchies

Deeply nested groups create visual clutter when fully expanded. Collapse everything you’re not actively examining. A collapsed group still surfaces its most critical status through bubbling, so you won’t miss problems.

The Tree panel is especially useful here. It shows the full hierarchy structurally, independent of how the graph is laid out spatially. You can navigate to any node by its position in the tree, which is often faster than panning around looking for it on the canvas.

Search is also a fast way to navigate. Open search with Ctrl+F (Windows/Linux) or Cmd+F (Mac), type a query, then double-click a result. You can also focus a result row and press Enter. The viewer reveals the node’s parent groups and focuses the camera on it.

Search results also split into visible and hidden nodes. If you’re looking for something and it doesn’t appear in the visible results, it’s probably filtered out. The hidden results tell you that immediately, and you can add those nodes back into the view directly from the search panel.

Combining filters progressively

Filters are stackable. Rather than trying to build the perfect filter in one step, work iteratively:

Isolate a broad area (e.g. all nodes of a certain type).
Exclude the noise you don’t care about.
Expand relations on a specific node to pull in its connections.
Bookmark the result.

Each filter narrows or expands the view incrementally. The Filter panel shows your full stack, so you can always see what’s shaping the current view and peel back layers if you’ve gone too far.

Keeping your bearings

A few habits help when you’re deep in a filtered or zoomed-in view:

Centre view often. Ctrl+Space / Cmd+Space reframes everything visible. Use it after applying or removing filters to re-orient.
Name your bookmarks well. “Interesting view 3” is useless next week. “Payment service dependencies” tells you exactly what you’ll get.
Use the Tree panel as a map. Even when the canvas view is heavily filtered, the Tree panel can show the full hierarchy for context.

Making the most of bookmarks

Bookmarks capture your current filters and group expansion state as a named snapshot. They’re stored per-document, so each graph has its own set.

A few ways to use them effectively:

Task-based views. Keep separate local bookmarks for recurring tasks. “Deployment topology” and “Service layer” can restore different filters over the same graph.
Investigation checkpoints. When troubleshooting, bookmark your starting point before applying aggressive filters. If you go down the wrong path, restore the bookmark instead of manually undoing filters.
Before/after comparisons. Bookmark a view, change your filters or group expansion, then switch back and forth to compare.

Bookmarks are browser-local. They do not sync to another browser, device, or team member.

Bookmarks can be renamed, updated with the current state (overwriting the previous snapshot), or deleted. All management happens in the Bookmarks panel.

What’s next