Reference scenarioSplunk Observability CloudOpenTelemetryTechnical · 3 min read
The service map shows “unknown_service”, or the same application five times.
Written by Fernando Mateo Ruiz · Splunk Enterprise Certified Architect · Splunk O11y Cloud Certified Metrics User · Reviewed on 28 October 2026
Reference scenario: it describes a common situation and how we tackle it. It is not a client project and includes no measured results.
In 30 seconds
The problem: After instrumenting, the service map shows “unknown_service”, the same application under several names, or production mixed with test.
What solves it: A resource attribute convention defined before deployment and applied in a single place.
What changes: The map reflects the real architecture, environment filters work and detectors don't break.
Sound familiar?
- The map shows
unknown_service:java. payments,payment-svcandPaymentsappear as three different services.- Test data shows up on production dashboards.
- Each team instruments in its own way.
The question to answer
What is each service called, which environment and version does it run in, and is it the same across all signals?
Why it matters
Without consistent names there's no reliable service map and no correlation between signals, and detectors and SLOs are built on exactly those names.
Signals needed
- Essentialservice.name
- Essentialdeployment.environment
- Recommendedservice.version
- RecommendedHost attributes
Architecture
Reference scenario: the architecture is generic and does not belong to any client.
How it’s implemented
- Write the convention down: the
service.nameformat (for example, lower case with hyphens) and the allowed values ofdeployment.environment(prod, pre, test). - Set the attributes in each service with environment variables, ideally in the shared deployment template.
- Add a safety net in the Collector: the
resourceprocessor withinsertonly adds the attribute when it's missing. - Review the map and the service list after every deployment.
OTEL_SERVICE_NAME=payments
OTEL_RESOURCE_ATTRIBUTES=deployment.environment=prod,service.version=1.4.2
processors:
resource/environment:
attributes:
- key: deployment.environment
value: prod
action: insert
# Add it to the processors list of your pipelines.
Step-by-step investigation
An illustrative example of how it would be investigated, from signal to action.
- 01 Signal
unknown_service:javaappears on the map with heavy traffic. - 02 ContextIt's a new deployment that didn't set
OTEL_SERVICE_NAME. - 03 CorrelationIts spans have the same hosts and endpoints as the
ordersservice. - 04 CauseA replica of
orderswas deployed with a different template. - 05 DecisionFix it in the deployment template, not by hand in the service.
- 06 ActionAdd the attributes to the shared template and a check that rejects deployments without
service.name.
Common mistakes
- Leaving the default name.
- Putting the environment in the service name (
payments-prod) instead of usingdeployment.environment. - Renaming services later and breaking detectors and dashboards.
- Using high-cardinality values as resource attributes.
What this case doesn’t solve
- It doesn't fix data that has already been sent.
- It doesn't replace an inventory of services and their architecture.
Expected result
Each service appears once, with its environment and version, and environment filters keep production apart from test.
Observed result
This is a reference scenario: we don’t publish measured results until it has been tested in our lab or in an authorised project.
Requirements
Platform
Splunk Observability Cloud
Access
Deployment configuration (variables, templates, Helm) and the Collector
Agreement
A shared convention across teams
Related services
Questions we’d ask you
- Do you have a written convention for naming services?
- How many teams instrument applications?
- Do you use shared deployment templates?