Skip to content

Limesium banner

Limesium

Limesium logs one structured endpoint_* line per HTTP exchange at the service's own boundary — named after the Roman Limes, the watched frontier where every crossing was recorded. Two auto-configured Spring Boot twins with identical fields and identical configuration: a servlet filter and a WebFlux/coroutines web filter. No starter, no forced transitives.

Module Stack Root package
limesium-servlet-logging Spring MVC / servlet filter eu.inqudium.limesium.servlet.logging
limesium-reactive-logging Spring WebFlux (Reactor and coroutines) eu.inqudium.limesium.reactive.logging

Features

  • Exactly one line per exchange. Sync and async share a single, exactly-once emission path; the outcome (success / failure / timeout) is decoupled from the log level.
  • A stable field contract. The endpoint_* wire names are a contract with the log index: each field owns its JSON shape, a badly typed value drops that field with a warning but never the event, and the Elasticsearch component template ships with the project — kept in lockstep with the code by contract tests.
  • Correlation built in. The correlation id is adopted from the configured header (or generated), echoed on the response, and rides the MDC for the whole chain — previous MDC values are restored afterwards.
  • Passive body capture, logged by outcome. Bodies are captured by a bounded tee as they flow — nothing is replayed or withheld from the application — and logged never, on-failure or always per direction; on-failure keeps the volume at the lines a body is wanted for. Logged header values are masked by default to a stable length:hash fingerprint (keyed on request, or whatever a host-provided HeaderValueMasker bean renders), plaintext being an explicit allowlist.
  • Twin symmetry as an invariant. Both modules expose the same fields and the same endpoint-logging.* properties; the shared reference configuration is contract-tested against both twins.
  • A library, not a platform. Auto-configured Spring Boot modules with no starter and no forced logging transitives; the host application brings the runtime (Tomcat resp. Netty) and the Logback binding.

Quick start

limesium-servlet-logging on Maven Central limesium-reactive-logging on Maven Central

The badges show the current release on Maven Central — use that version where the snippets say .... Add the module matching your stack — the filter registers itself:

<dependency>
    <groupId>eu.inqudium</groupId>
    <artifactId>limesium-servlet-logging</artifactId>
    <version>...</version>
</dependency>

or, for a WebFlux application:

<dependency>
    <groupId>eu.inqudium</groupId>
    <artifactId>limesium-reactive-logging</artifactId>
    <version>...</version>
</dependency>

Every endpoint-logging.* key, with its default, is documented in the configuration reference — copy the block and change only what you need.

Documentation

  • Common guide — everything both modules share: the exchange line, the shared architecture, dependency and encoder setup, the configuration namespace, the field family, the meters, the trace contract, and the table of deliberate stack differences.
  • Servlet guide — what the servlet stack decides in the reference implementation: the filter and its two registrations, request destruction as the emission point, async exchanges, the chain-wide MDC, the servlet-only edge cases.
  • Reactive guide — what the reactive stack decides: the two filter variants, the commit-deferred emission, the Reactor context and handler-side MDC, the reactive-only edge cases.
  • Elasticsearch mapping — the ready-made component template for the endpoint_* fields.
  • Test evidence — the generated inventory of the test suite: every test sentence plus its rationale, grouped by module and component.
  • Coverage report — the JaCoCo reports of the run that built this site.
  • API reference — servlet and reactive, generated with Dokka.

Project