Precision of timestampNanos in logs is miliseconds instead of nanoseconds

[rafal-dudek]

Describe the bug
StackdriverJsonLayout sets value of 'timestampNanos' log field as
TimeUnit.MILLISECONDS.toNanos(event.getTimeStamp() % 1_000) which has precision of miliseconds.
https://github.com/GoogleCloudPlatform/spring-cloud-gcp/blob/main/spring-cloud-gcp-logging/src/main/java/com/google/cloud/spring/logging/StackdriverJsonLayout.java#L276
Proper value should be event.getNanoseconds()

This is causing out-of-order logs in GCP Logs Explorer in some cases.

Sample

    Logger log = LoggerFactory.getLogger(GenericSpringBootApplication.class);
    log.info("1");
    log.info("2");
    log.info("3");
    log.info("4");
    log.info("5");

Result in kubectl logs:

{"timestampSeconds":1688547662,"timestampNanos":662000000,"severity":"INFO","thread":"scheduling-1","message":"1"}
{"timestampSeconds":1688547662,"timestampNanos":663000000,"severity":"INFO","thread":"scheduling-1","message":"2"}
{"timestampSeconds":1688547662,"timestampNanos":663000000,"severity":"INFO","thread":"scheduling-1","message":"3"}
{"timestampSeconds":1688547662,"timestampNanos":663000000,"severity":"INFO","thread":"scheduling-1","message":"4"}
{"timestampSeconds":1688547662,"timestampNanos":663000000,"severity":"INFO","thread":"scheduling-1","message":"5"}

Result in GCP Logs Explorer:

[meltsufin]

Would you mind submitted a fix?