Logging.md

HMCTS API Marketplace Logging

Example Tests

See JunitLoggingTest.java See SpringLoggingIntegrationTest.java These tests illustrates the fields that we add to our logging

Using logback.xml

Configuring logging with logback.xml is the simplest approach We can add any global logging fields by adding to MDC in any request filter i.e. In TracingFilter.java we have these lines to add applicationName to every log entry

        public TracingFilter(@Value("${spring.application.name}") String applicationName)
        ...
        MDC.put(APPLICATION_NAME, applicationName);

Using logback-spring.xml

Currently, configuring logging with logback-spring.xml is more complex and seems to be quite fragile. Bringing in dependencies that include logback.xml or commons-logging can cause the logging to revert to basic logging It is imperative that any usage of logback spring is covered with a strong integration test such as the example SpringLoggingIntegrationTest.java

Collecting logs in Azure

We want to produce logging as json with fields as detailed below We currently output to stdout and expect that the k8s containers will be wired to capture the logging into azure logging Once we work closer with Azure ops teams we expect to implement an spring boot azure plugin to directly inject logs to azure

Fields

See logback.xml for master list. Or better still, run the unit test or integration to see the actual logged entries. mdc - Logs any objects added to MDC e.g. io.micrometer adds traceId and spanId to MDC. We can add any custom fields to this.

Testing

We have strong unit and spring boot integration tests that assert the individual fields in log entries We dont currently have a test to prove the main Application logging is correct This may need to be done outside the app test suite maybe as part of a docker test