Open Stack

Hi All,

I am announcing my candidacy for PTL of the OpenStack Vitrage project for the
Pike cycle.

I have been the Vitrage PTL from its first day. I was involved in the first
stages of the design and implementation (in Mitaka), was thrilled when it
became an official OpenStack project only seven months later (in Newton),
and spent more and more time on relationships with our growing community (in
Ocata).

I think that the Vitrage project has a group of extremely talented developers,
who achieved in less than a year and a half all of the major goals we set for
ourselves. Seeing how Vitrage grew from an idea to a mature, production-grade
and well-known project in such a short period of time was an amazing experience
for me.

With that said, we have quite a few challenges ahead. I’ll describe the areas
that I believe we should focus on in the Pike cycle.

- Extend our community. As time passes, we see more and more interest in
  Vitrage. The more contributors we have, the better Vitrage can get.

- Support more use cases
  - Deduce alarms that are not reported as expected (lost or delayed)
  - Add more deduced alarms and RCA templates (e.g. for network monitoring).
    It took us a while to build the infrastructure, but now it’s there.
    It’s time to think of new alarms, what effect they will have on the system,
    and who can be notified and benefit from this information.

- Integrate with more OpenStack and external projects, for the sake of the
  above goal.

- Improve Vitrage usability. Vitrage provides a lot of valuable information
  that is presented in the Horizon UI. As much informative as it is, the way
  it is displayed is not ideal and should be enhanced.

- Support a persistent graph database. We have been talking about it for a
  while, it’s time to implement. Our in-memory graph database works very well,
  but a persistent one has its own advantages.

- Enhance the Vitrage evaluator templates language, and support full template
  CRUD API.

Overall, we would like Vitrage to become a project that every cloud operator
would like to use, and I believe we are on the right direction. I think that
the Pike cycle will be a very interesting one.

Thanks,
Ifat.