Do I need service maps from traces or is that overkill

It depends on what you’re trying to do, but in many cases service maps from traces are useful, not overkill.

You probably want service maps if:

• You have multiple services and want to see how they interact
• You’re trying to find where latency or errors originate
• You need to understand call paths between services
• You’re doing incident response or troubleshooting intermittent issues
• You want a quick topology view of dependencies

It may be overkill if:

• You only have one or two services
• Your system is simple and stable
• You already know the dependency graph well
• You mainly care about metrics/logs and traces are only for deep debugging
• You don’t have enough trace coverage to make the map reliable

A practical rule

• Small/simple system: start with traces + metrics, skip service maps initially
• Microservices / many dependencies: service maps are very helpful
• If you’re unsure: enable them if your tracing backend supports it with minimal extra effort

Important caveat

A service map is only as good as your trace coverage. If you’re missing spans, sampling heavily, or instrumenting inconsistently, the map can be misleading.

If you want, I can help you decide based on your architecture — e.g. “single app,” “few services,” or “microservices on Kubernetes.”