Thank you!

[Mark Shuttleworth]

On 20/04/17 13:27, Merlijn Sebrechts wrote:
> What about speed an velocity? Does a more focused Canonical mean more
> resources for Juju?

With good reason the team is unlikely to be comfortable answering your
question, but I am happy to do so.

Juju is essential to  our success with large customers every day. It
makes it possible for us to support complex OpenStack and Kubernetes
deployments that are regularly extended and upgraded and updated,
something that nobody else can in practice deliver. And those customers
are increasingly wanting to use it for a broader array of software. So
Juju has its place in our plan and I am very happy with the way the team
is now organised, with team leads who really care about quality,
performance, features and customers. I hope you've noticed a big step up
in the quality and focus of Juju releases since 2.x and I would credit
the teams and their leads with that.

I asked every single team at Canonical, with a few specific exceptions
for example LXD and the security team, to participate in the cut, on the
basis that I wanted every manager to have to exercise this kind of
judgment. It's tough, but it's also important for a company to be
confident that it is constantly improving and assessing itself.

In the case of Juju Core and JAAS, I had already driven some changes in
this regard, and I felt the team was already substantially more focused
and doing well, so there was a gentler approach, although it is always
an emotional and difficult process even if it touches relatively few
colleagues. We will miss those colleagues but I think you can count on
steady and exciting progress in Juju and JAAS.

In the teams who work on charms, we are concentrating more folks around
the key pillars of OpenStack and Kubernetes, which are manifestations of
the "os-at-scale" so we (Canonical) are the natural vendor. We are also
asking the folks who lead charm-tech to use the K8s charms as their
proving ground, to ensure those tools really work for complex charm
development. We will also continue to anchor the Hadoop ecosystem but I
am delighted with the growth in third-party interest in charms in that
ecosystem and I am happily counting on you, Merlijn, and the other
members of the Hadoop charm ecosystem, to drive forward the broader
suite of charms there and elsewhere at the application level.

Mark