Would you use StatLite for non-Spring Java applications?

[ejboy]

StatLite currently uses Spring Boot Actuator as its primary Java integration. That works well for Spring Boot applications because it requires no application changes: StatLite polls a focused set of existing Actuator endpoints and stores the resulting history locally.

I'm considering whether it would be useful to support regular Java web applications that do not use Spring Boot.

One possible approach would be a very small Java integration with two pieces:

• a lightweight Servlet filter that collects HTTP request metrics such as request count, 4xx/5xx responses, active requests, and request latency;
• a small endpoint that exposes JVM/process and application metrics using StatLite's fixed statlite-metrics/v1 format.

JVM-level metrics such as memory and process CPU could come directly from standard Java management APIs, while the filter would provide the request-level information that is otherwise application-specific.

The goal would not be to build another general metrics library or Prometheus exporter. I would want to keep the integration narrowly scoped to the metrics StatLite actually displays and preserve the same philosophy as the rest of the project: lightweight, well-tested, predictable, and easy to remove.

A Servlet-based integration could potentially cover applications running on Tomcat, Jetty, Jersey, and similar stacks. It would not automatically cover every Java server or reactive framework, so I would rather add adapters based on real demand than try to create a universal abstraction upfront.

Before building this, I'd like to understand whether there is an actual use case for it.

If you would use StatLite with a non-Spring Java application:

• What framework or server are you using?
• Is it Servlet-based, or something else such as Netty, Vert.x, Undertow, etc.?
• Would adding a small StatLite library/filter dependency to the application be acceptable?
• Which metrics would actually matter to you?
• Would a focused built-in endpoint be preferable to configuring a more general metrics stack?

Or is Spring Boot Actuator coverage already enough for the kinds of small deployments where StatLite makes sense?

I'm especially interested in concrete applications or deployment setups rather than adding broader Java support speculatively.

[ejboy]

My current thought is to start with a small copyable Java example, similar to the existing Python integration, rather than introducing a new library immediately. If a common pattern emerges across real applications, it could later be packaged as a small reusable dependency.