回退消息在多线程dropwizard服务方案中变得混乱

Question

我们有一个多线程服务，当另一个线程同时写入日志时，某些单独的日志消息会被破坏。我怎样才能防止这种情况发生？

我们正在使用工作线程来进行REST服务调用，接收和处理响应。在执行此操作时，我们使用logback向dropwizard中的默认日志记录引擎写入日志。

我正在看的具体情况是当线程抛出异常时，我们记录堆栈跟踪。我看到的是，跟踪被至少一个其他线程的条目减少了一半：

实施例

   [pool-4-thread-13] Error!
         stack trace line 1
   [pool-4-thread-17] Event 1 happened
   [pool-4-thread-17] Event 2 happened
         stack trace line 2
         stack trace line 3
         stack trace line 4

我期望在日志中看到的内容是这样的：

   [pool-4-thread-13] Error!
         stack trace line 1
         stack trace line 2
         stack trace line 3
         stack trace line 4
   [pool-4-thread-17] Event 1 happened
   [pool-4-thread-17] Event 2 happened

config.yaml

server:

  adminContextPath: /admin
  adminConnectors:
    - type: http
      port: 9193

  applicationContextPath: /

  applicationConnectors:
    - type: http
      port: 9192
      maxRequestHeaderSize: 64KiB

  requestLog:
    timeZone: UTC

    appenders:
      - type: file
        currentLogFilename: ./logs/service-requests.log
        archive: true

        # When the log file rotates, the archived log will be renamed to this and gzipped. The
        # %d is replaced with the previous day (yyyy-MM-dd). Custom rolling windows can be created
        # by passing a SimpleDateFormat-compatible format as an argument: "%d{yyyy-MM-dd-hh}".
        archivedLogFilenamePattern: ./logs/service-requests.log.%d.log.gz

        # The number of archived files to keep.
        archivedFileCount: 30

logging:
  level: INFO

  loggers:
    # We only care about warning log messages from the apache libraries, like hadoop, zookeeper, and phoenix
    org.apache: WARN

  appenders:
    - type: file
      timeZone: UTC
      currentLogFilename: ./logs/service-log.log
      archive: true
      archivedFileCount: 30
      archivedLogFilenamePattern: ./logs/service-log-%d.log.gz
      logFormat: "[%date{dd MMM yyyy HH:mm:ss}] [%thread] %-5level %logger - %message %n"