It's just come to my attention that the Newton release notes indicated deprecation of the events support in favor of Panko [1] and that it has already been removed in master [2]. But I don't believe there were any deprecation warnings for this in the Newton code, which is the typical way many folks learn about deprecations. The only deprecation warning I've found was merged after Newton released [3]. And while the official OpenStack deprecation policy allows you to remove things in the next release after deprecation, it also says "Note that this delay is a required minimum. For significant features, it is recommended that the deprecated feature appears at least in the next two stable release branches." [4]. So a) we should have had a deprecation warning in the code to go along with the releasenote in Newton and b) even then we probably shouldn't be removing the event API support until Pike. I started the conversation on IRC [5], but wanted to send this to the mailing list and see if others have thoughts/concerns here and figure out what we should do about this going forward. [1] http://docs.openstack.org/releasenotes/ceilometer/newton.html#deprecation-notes [2] https://github.com/openstack/ceilometer/commit/8d23f431ab0bd638edbf2197e56bea68d7b06a21 [3] https://github.com/openstack/ceilometer/commit/6616a714009a80c7484fa2292c2331868617cb9c [4] https://governance.openstack.org/tc/reference/tags/assert_follows-standard-deprecation.html [5] http://eavesdrop.openstack.org/irclogs/%23openstack-telemetry/%23openstack-telemetry.2017-01-09.log.html#t2017-01-09T17:24:36 -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20170109/7250ee92/attachment.html>