Service Map
Service Map gives you a real-time, visual representation of your service architecture based on your distributed traces. It automatically discovers services and their dependencies from OpenTelemetry trace data, showing you how requests flow through your system along with key performance indicators for each service.
Instead of digging through individual traces to understand your architecture, Service Map shows you the full picture - which services talk to each other, how they're performing, and where problems are occurring.
Note: Service Map is currently available as a beta feature.
Why Use Service Map?¶
Traditional monitoring tells you that a service is slow or throwing errors. Service Map shows you the bigger picture - which services are affected, what they depend on, and where the problem originates.
With Service Map, you can:
See your architecture as it actually is. Service Map is built from real trace data, not static diagrams or manual configuration. Every service and dependency you see reflects actual request flow, so you're always looking at the truth.
Investigate incidents faster. When a service starts failing, open Service Map to immediately see its upstream callers and downstream dependencies. Spot cascading failures, identify the root cause service, and jump straight into filtered traces with one click.
Track service health at a glance. Color-coded nodes and edges surface degraded or critical services without digging through dashboards. Error rates above your thresholds turn nodes red, and high-latency dependencies stand out with orange or red edges.
Understand latency patterns. Each service node shows P95 and P99 latency, and the details panel includes a latency distribution histogram. Quickly tell whether slowness is systemic or caused by occasional outliers.
Validate deployments and changes. After deploying a new service or modifying communication patterns, use Service Map to verify dependencies are correct and no unexpected connections have appeared.
Onboard faster. New team members can open Service Map to understand how services interact without reading outdated documentation or tracing through code.
Accessing Service Map¶
There are several ways to access Service Map:
- From the top-level navigation: Click Service Map under Insights and Analysis in the left menu, select a Tracing App, and click Go to Service Map
- From within a Tracing App: Open any Tracing App and click Service Map in the left sidebar
- From Logs or Synthetics Apps: Navigate to Service Map from within your Logs or Synthetics Apps to quickly visualize the service dependencies related to your troubleshooting context
Being accessible from multiple entry points - Tracing, Logs, and Synthetics - means you can jump to Service Map during any troubleshooting workflow without losing context.
The map is generated from trace data flowing into your Tracing App, so it reflects actual service communication patterns - not static configuration.
Understanding the Map¶
Nodes (Services)¶
Each node on the map represents a service detected in your traces. Nodes display key metrics at a glance:
- Service Name: The name reported by your OpenTelemetry instrumentation
- Role Badge: Whether the service acts as a Server or Client in the traced interactions
- Request Rate: Number of requests per second handled by the service
- P95 Latency: 95th percentile response time
- P99 Latency: 99th percentile response time
- Error Rate: Percentage of requests resulting in errors
Node Health Status¶
Nodes are color-coded based on their error rate:
| Status | Color | Default Threshold |
|---|---|---|
| Healthy | Green | < 2% errors |
| Degraded | Orange | 2-4% errors |
| Critical | Red | > 4% errors |
| No Data | Gray | No recent trace data |
Edges (Dependencies)¶
Lines connecting nodes represent service-to-service dependencies discovered from trace data. Edge appearance indicates the health of the communication:
| Status | Appearance | Default Threshold |
|---|---|---|
| Healthy | Green solid line | < 500ms latency |
| High Latency | Orange solid line | >= 500ms latency |
| Errors Present | Red animated dashed line | Errors detected |
The Quick Guide button in the toolbar provides a handy reference for these visual indicators.
Node Details Panel¶
Clicking on any node opens a details panel with deeper insights into that service.
Quick Actions¶
Jump directly to filtered trace views:
- View All Traces: Opens Traces Explorer filtered to this service
- View Slow Traces: Opens Traces Explorer filtered to high-latency traces for this service
- View Error Traces: Opens Traces Explorer filtered to traces with errors for this service
Overview¶
Expanded metrics for the selected service:
- Request Count
- Error Rate
- Avg Latency
- P95 Latency
- P99 Latency
Latency Distribution¶
A histogram showing how request latencies are distributed, helping you understand whether latency issues affect all requests or just a tail.
Dependencies¶
Lists the downstream services this node communicates with. You can filter dependencies by direction (All, Incoming, Outgoing) and sort by Requests, Response Time, or Error Rate. Each dependency entry shows:
- Request rate
- Average response time
- Error rate
Top Endpoints¶
Shows the most frequently called endpoints for this service with their request counts. This is valuable for understanding which endpoints are actively being used - helping you identify unused endpoints, discover undocumented APIs, and verify that traffic is flowing where you expect it.
Edge Details Panel¶
Clicking on any edge (connection line) between two services opens a details panel showing the health and characteristics of that dependency.
The panel header shows the source and target services with their key metrics (request rate, error rate, P95 latency) and the dependency direction.
Quick Actions¶
- View Traces (Both): Opens Traces Explorer showing traces that involve both services
- View Slow Requests: Shows only high-latency traces between the two services
- View Failed Requests: Shows only traces with errors between the two services
Request Metrics¶
- Total Requests: Number of requests between the two services
- Successful: Count and percentage of successful requests
- Errors: Count and percentage of failed requests
Latency Metrics¶
- Avg Latency: Mean response time for requests between the services
- P50 Latency: Median response time
- P95 Latency: 95th percentile response time
- P99 Latency: 99th percentile response time
Protocol Information¶
Shows the communication protocol (e.g., HTTP) and the breakdown of primary methods used (e.g., GET, POST) with their percentages.
Top Endpoints¶
Lists the most frequently called endpoints in this service-to-service communication.
Filtering the Map¶
Click the Filters button in the toolbar to open the filters panel, which lets you focus the map on specific services or conditions.
Health Status¶
Filter nodes by their health status: Healthy, Degraded, Critical, or No Data.
Service Kind¶
Filter by service role: Server or Client.
Performance Metrics¶
Set range filters on key metrics to highlight services matching specific criteria:
- Error Rate Range: Show only services within a specific error rate range (e.g., 0-5%)
- Latency Range: Show only services within a specific latency range
Topology Filter¶
Focus on specific parts of the service graph:
- Only services with errors: Show only services that have errors
- Only services with high latency: Show only services exceeding the latency threshold
Customizing Thresholds¶
Default health thresholds may not fit every environment. Click the Thresholds button in the toolbar to customize warning and critical levels for:
Service Error Rate¶
Configure what error rate percentage is considered healthy, warning, or critical for node coloring.
Latency Threshold¶
Configure the latency value (in milliseconds) that determines whether an edge is shown as healthy or high-latency.
Changes apply immediately to the map visualization.
Toolbar Controls¶
The toolbar at the top of the Service Map provides:
- Targets: Filter the map to show specific services or service groups
- Thresholds: Customize health indicator thresholds
- Filters: Apply additional filtering criteria
- Quick Guide: Reference card for node and edge visual indicators
- Pause/Play: Pause or resume real-time data updates
Service Map vs. Network Map¶
Sematext offers two complementary topology visualization features:
| Service Map | Network Map | |
|---|---|---|
| Data Source | OpenTelemetry traces | eBPF network probes |
| Scope | Application-level service dependencies | Infrastructure-level network connections |
| Metrics | Request rate, latency, error rate | Traffic bytes, CPU, memory, I/O |
| Accessed From | Tracing App or top-level nav | Infra App |
| Best For | Understanding request flow and service health | Understanding infrastructure topology and network traffic |
Use Service Map when you need to understand how your application services interact and where performance bottlenecks exist. Use Network Map when you need to understand infrastructure-level connectivity, resource consumption, and network traffic patterns.
Next Steps¶
- Search and analyze traces for specific services
- View detailed trace waterfalls to understand individual requests
- Set up alerts on latency and error rates
- Explore Network Map for infrastructure-level topology