Open Stack

Hi,

Barcelona summit was very successful and productive for the Vitrage team. Below is a summary of our main activities during the summit.

Sessions
--------

1. Keynotes demo - OpenStack and OPNFV - Keeping Your Mobile Phone Calls Connected

Ifat Afek, Ryota Mibu and Ildiko Vancsa presented how OPNFV Doctor can be used to prevent a phone call from being disconnected, and how Vitrage can be used to visualize the problem that occurred and the affect it had on the system.
https://www.openstack.org/videos/video/demo-openstack-and-opnfv-keeping-your-mobile-phone-calls-connected

2. Fault Management with OpenStack Congress and Vitrage Based on OPNFV Doctor Framework

Ohad Shamir, Ryota Mibu and Masahito Muroi presented the OPNFV Doctor requirements for fault management framework for NFV, and how Vitrage and Congress can help meet these requirements.
https://www.openstack.org/videos/video/fault-management-with-openstack-congress-and-vitrage-based-on-opnfv-doctor-framework

3. Nokia Root Cause Analysis Principles and Practice in OpenStack and Beyond

Elisha Rosenzweig and Ifat Afek presented the different concepts for root cause analysis, and how the Vitrage project can be used for organizing, visualizing and expanding OpenStack events and alarms.
https://www.openstack.org/videos/video/nokia-root-cause-analysis-principles-and-practice-in-openstack-and-beyond 

4. Vitrage Hands-On Lab

Dan Offek presented Vitrage and guided the participants throughout the process of installing, configuring and experimenting with Vitrage on their own laptops. Over 40 people participated in this lab.


Design Sessions
---------------

An etherpad with the summary of all Vitrage design sessions: https://etherpad.openstack.org/p/vitrage-barcelona-design-summit

1. Vitrage fishbowl session - Taking the Vitrage UX to the Next Level

This design session was dedicated to discussing ways of improving Vitrage usability, mostly around the entity graph UI. Some ideas that were raised are: 

- Improve the search capabilities 
- Show different views for different personas
- Show a high level aggregation and then drill down
- Show only the “red” areas of the graph
- Use a non-Horizon UI for the graph visualization

Another issue that was discussed was how to represent alarms history in the UI.


2. Vitrage Work Session

The first work session focused on technical challenges in Vitrage: 

- How to support RCA history. We discussed the need to store the alarms in a persistent (graph?) database, and the performance issues it might raised. We also talked about logical questions regarding how to identify or display different root causes that change over time.

- Vitrage Log evaluator. A suggested offline tool to analyze Vitrage logs and explain what happened. It could be useful when trying to figure out why a certain template was not executed for example.

- Performance benchmark. We discussed different methods of doing performance benchmark of the Vitrage evaluator.


3. Vitrage Work Session 

The second work session was focused on Vitrage integrations with other projects:

- Vitrage-Aodh integration. Vitrage already polls alarms from Aodh in Newton. We discussed the implementation of sending notifications from Aodh to Vitrage whenever an alarm state is changed. In addition, we discussed how Vitrage alarms can be raised in Aodh, so Aodh users will be able see Vitrage alarms in Aodh. It was agreed that this requires adding a new alarm type in Aodh, and that this task will be performed by someone from Vitrage team.

- There was a question whether Vitrage can benefit from an integration with Panko or Gnocchi. This requires a further research.

- Vitrage-Congress integration. We discussed the option of creating policies based on root cause analysis insights that Congress receives from Vitrage. The same alarm may be handled differently by Congress, based on its root cause. We agreed that we should identify concrete use cases, and talked about some implement details.


Best Regards,
Ifat.