Andrea Cosentino created CAMEL-24498:
----------------------------------------
Summary: camel-observability-services-starter - injected
management defaults exceed the Spring Boot baseline
Key: CAMEL-24498
URL: https://issues.apache.org/jira/browse/CAMEL-24498
Project: Camel
Issue Type: Improvement
Components: camel-spring-boot-starters
Reporter: Andrea Cosentino
Assignee: Andrea Cosentino
{{ObservabilityServicesEnvironmentPostProcessor}} injects a set of management
defaults whenever the starter is on the classpath:
{code:java}
defaults.put("management.server.port", "9876");
defaults.put("management.endpoints.web.exposure.include", "health,prometheus");
...
defaults.put("camel.health.exposure-level", "full");
defaults.put("management.endpoint.health.show-details", "always");
defaults.put("management.endpoint.health.group.live.show-details", "always");
defaults.put("management.endpoint.health.group.ready.show-details", "always");
{code}
(ObservabilityServicesEnvironmentPostProcessor:40-60)
Three of these are broader than the Spring Boot defaults they replace:
- {{management.server.port}} is set without a matching
{{management.server.address}}, so the second listener binds all interfaces.
Spring Boot has no default management port at all.
- {{management.endpoint.health.show-details}} is {{always}} where Spring Boot's
default is {{never}}.
- {{camel.health.exposure-level}} is {{full}} where the Camel default is
{{default}}.
The property source is added with {{addLast}}, so user configuration wins -
this is about what an operator gets when they add the starter and configure
nothing.
*Proposal*
- Set {{management.server.address=127.0.0.1}} alongside the port so exposing
the listener is a conscious step.
- Default {{show-details}} to {{when-authorized}}, keeping {{always}} for the
{{live}}/{{ready}} probe groups if the k8s probe flow needs it.
- Leave {{camel.health.exposure-level}} at its default and document {{full}} as
an opt-in.
- Document the combination in the starter's docs.
Note for the release notes: Kubernetes deployments whose kubelet probes or
Prometheus scrapers reach the pod over the network would need to set
{{management.server.address=0.0.0.0}} explicitly after this change. That is a
deliberate behaviour change and should be called out in the upgrade guide.
----
_This issue was drafted by Claude Code on behalf of Andrea Cosentino._
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
