Hi everyone, Some offline conversations and thoughts yielded a notion of providing a formal policy/transition backing to on-error, on-success and on-finish. I have a few thoughts in and etherpad (see https://etherpad.openstack.org/p/mistral_task_policy). The headliner : on-error, on-success and on-finish should be more complex than defining what task to call next. Look forward to the comments. Thanks, Manas -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20140410/73a6bdec/attachment.html>