RFE: Backend operational metric compatibility and migration guidance #6001

The backend already exposes HTTP/runtime metrics, management endpoints, Prometheus output, and OpenTelemetry export. Existing dashboards and integrations that depend on Micrometer/Actuator or JDBC-pool meter names cannot be moved over simply by switching imports. Some JVM-only measurements also have no direct native equivalent.

Requested capability

• Publish a stable metric schema with names, units, labels, aggregation, and lifecycle semantics for HTTP, database pools, executors, memory, and scheduling.
• Provide an explicit mapping guide or optional adapter for common Micrometer/Actuator use cases.
• Expose missing equivalent measurements where feasible and distinguish unavailable metrics from measured zero.
• Document readiness, liveness, drain/shutdown, endpoint authentication, and dashboard migration behavior.

Acceptance criteria

• A sample dashboard works against JVM and native backends using the documented schema.
• Contract fixtures verify metric units, counters/timers, pool utilization, and lifecycle changes.
• Unavailable JVM-specific metrics are identified explicitly instead of silently returning zero.
• Existing Prometheus/OTLP exporters and management security remain supported.

Current implementation

• docs/developer-guide/Backend-Observability.asciidoc
• vm/backend/src/com/codename1/backend/HttpServer.java

Validate behavior on the JVM and locally on the packaged ParparVM backend where applicable. This request does not call for native-image or GraalVM CI jobs.