admin task threading

[Martin Edler]

We are having a strange issue when we run dropwizard admin tasks via the administrative HTTP port.

When the admin-task takes more than 3 minutes, maybe because there is lot of database stuff to do, after exactly 3 minutes an unexpected second thread starts to work. We see the threads in the in our logfiles, because the thread-name is part of the log messages. The first thread is called e.g. dw-amdin-77, the second one appears after 3 minutes under a similar name, e.g. dw-admin-137. From that point on, both threads emit log messages alternatively. Both threads seems to do the same work. Even if the admin-task is running for ours, we have never seen a third thread.

We are not aware of doing any asynchronous stuff. Our Dropwizard application runs in a docker container. The task is started manually by a single http-request.

Any ideas about what could cause this behaviour are welcome.

Thanks in advance,

Martin Edler

(dropwizard version 1.3.x)