Sometimes the forking model have desirable properties. Numerous past articles covered this in one way or another. Choosing between SideKiq and Resque is prima

Sometimes the forking model have desirable properties. Numerous past articles covered this in one way or another. Choosing between SideKiq and Resque is primarily a matter of choosing between forked processes and threads. Retiring Resque and let SideKiq be the present and the future would destroy that choice. SideKiq, as a threaded impl., can claim some cool benefits : - shorter job launch times (only interesting if you actually run thousands of second spanning jobs a minute) - reduced memory footprint (you probably have little to no amount of thread local storage) I will venture that the reduced memory footprint is mostly the result of the unfriendly to copy-on-write GC we have to live with using 1.9. Resque has a clean and simple API and a decent ecosystem. It can and hopefully will evolve into something cooler and more powerful - even if worthy alternatives exist.