Following the forum session yesterday I've thrown up this pseudocode for discussion: https://review.openstack.org/463618 I want to stress that I consider nova proxying to searchlight to be a premature optimisation rather than an intrinsically terrible idea. That said, the searchlight solution looks very complex for both developers and operators, and very fragile. I think we'd better going with a relatively simple solution like this one first, and only going a couple of orders of magnitude more complex if it turns out to be absolutely essential. Matt -- Matthew Booth Red Hat Engineering, Virtualisation Team Phone: +442070094448 (UK) -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20170509/aea8d0a5/attachment.html>