cloud-logging

Logging libraries for JVM applications. Goals

• make developers more productive when developing and maintaining applications
• keep cloud logging costs within reason
• help comply with privacy constraints for logging

Features:

• SLF4J Logger-wrapper for additional error levels
• Plug and play logging scheme
  • Main scope (for production):
    • Machine-readable JSON
  • Test scope (for local development). Log output can be toggeled at runtime:
    • Human-readable 'classic' (one-line + ANSI colors, enabled by default), or
    • Human-readable JSON (pretty-printed + ANSI colored), or
    • Machine-readable JSON (single line)
• Request-response-logging
  • Logbook style output (i.e. inlined in JSON log statements)
  • Additional ANSI coloring for test scope
• Selective 'on-demand' logging for unexpected web server behaviour
  • capture full logs for problematic requests (i.e. not only WARN or ERROR, but also all sibling INFO log statements)
  • reduce cost of logging considerably
• Unit testing
  • Always assert against machine-readable JSON 'under the hood', regardless what is printed to console during local development
  • Supported frameworks
    • JUnit 5
• Correlation-id tracing

This library is mostly based on

• SLF4J 2
• Logback + Logback logstash
• Lognet GRPC
• Logbook

Supported web technologies are:

• Spring web
• Spring + gRPC via Lognet
• Micrometer

Supported clouds:

• GCP (stackdriver)

License

European Union Public Licence v1.2.

Usage

Overview

This project contains Spring boot starters which allows for property-based configuration:

• base configuration (required)
• request-response-logging (optional)
• on-demand-logging (optional)

Each Spring Boot starter has a corresponding Spring Boot test starter. The Spring boot test starters enhance the developer experience during local development.

The above starters are implemented in two flavours:

• Servlet-based web
• Lognet gRPC

Getting started

See Getting started with gRPC or Getting started with servlet-based web. Alternatively, go directly to the examples.

Roadmap

• Replace the correlation-id tracing artifacts with standardized modern equivalents, i.e. span id and so on.
• Better define and tune request-response logging JSON format