* Introduce pull-timeout * Keep isolate shutdown and write\_file functions * Catch timeout during the connection
0.1.1
* Rework the retries decorator
0.1.0
* Shield workers from death due to errors from dispatchers
0.0.5
* Don't try to process empty Pandas data frame
0.0.4
* Break early when STOP item is fetched from queue * Don't enable local-store by default * Perform a clean shutdown in case no tasks are running * Wait for socket files to be created * Don't proceed if watched dir isn't created * Introduce a configuration parameter for process * Add systemd unit files