Service Dependency View

The Service Dependency view is the default Flow View map. It visualizes all services reporting to APM alongside their connected dependencies—including databases, message queues, caches, third-party APIs, and downstream services—allowing you to understand complete request flows and architectural dependencies without inspecting individual traces.

What is on the map

The view initializes with the following primary entity types:

Expanding a service

Double-click a service node to expand its transactions—the individual HTTP endpoints and background jobs it serves—rendered as child nodes within the service container.

When collapsed, a service displays an aggregated connection line to its dependencies (e.g., a single line to postgresql). When expanded, individual transactions reveal granular connections to the exact databases, message queues, or downstream services they invoke.

Expanding a Database system, Messaging system, or LLM provider similarly exposes the specific databases, topics, or AI models configured within that system.

Isolating a service

To focus on a specific service in a complex topology, select the service and click Isolate on its info card. Flow View filters the canvas to show only that service, its transactions, and its direct 1-hop upstream and downstream dependencies, removing unrelated entities from view.

Click Clear isolation at the bottom of the canvas to restore the complete topology.

Layouts

The Service Dependency view supports three layout modes:

• Vertical (default): Arranges services hierarchically with upstream callers above and downstream dependencies below. Click the active layout button to invert the flow direction.
• Horizontal: Arranges services with upstream callers on the left and downstream dependencies on the right. Click the active layout button to invert the flow direction.
• Force: Uses physics-based clustering to naturally position interconnected services and dependencies.

Reading the lines

In the Service Dependency view, lines represent application dependencies and communication paths:

• Selection highlight: Clicking any service or dependency illuminates its connected lines in animated blue and dims unrelated nodes.
• Line details: Click any line to open a details card displaying the source service, destination dependency, and relationship type.
• Uniform thickness: Line thickness is uniform across all connections; lines indicate that a dependency exists rather than request volume or throughput.

Troubleshooting workflow

When troubleshooting a degraded service (e.g., high response times on order_service):

1. Open the Service Dependency view and locate order_service using the canvas or search box (Ctrl/⌘+F).
2. Click the service to select it. Its immediate dependencies highlight while unrelated nodes dim, revealing the direct dependency path.
3. Check the health borders of connected dependencies. A red border indicates the root cause of an active Problem; an amber border indicates an open Issue.
4. Double-click order_service to expand its transactions and identify which specific endpoint communicates with the degraded dependency.
5. Click the degraded dependency and select Investigate problem on its info card to view incident diagnostics directly on the map.
6. Click through to APM from the info card to inspect detailed transaction traces, error rates, and metric breakdowns.

Next steps

• Navigating the graph: Canvas interactions, searching, and keyboard shortcuts.
• Problems View: Investigate active incidents directly from root-cause entities.