rubick/doc/openstack_integration.rst

75 lines
2.5 KiB
ReStructuredText

VALIDATOR INTEGRATION WITH OPENSTACK
====================================
--------
Overview
--------
OpenStack cloud life cycle consists of 2 distinct phases:
* initial deployment and
* operation maintenance
Success of the *initial deployment* stage means that upfront installation and
configuration of the OpenStack platform is successful. Multiple projects work in
that area: TripleO+Tuskar, Fuel, Packstack, Devstack to name a few.
Most of these projects are pretty good in what they are doing, but once you
installed and kicked off the cloud, things begin to change. Cloud enters
*operation maintenance* phase of its life cycle.
Changes to configuration, architecture and environment itself could reduce
overall stability and performance of the platform. Eventually operators faces
the situation when something goes wrong and service gets disrupted.
That said, *operation maintenance* phase is more or less abundant of
community-driven projects aimed to increase reliablility of cloud and
resilience which allows to support the SLA at reasonably high level (think your
favourite number of nines).
Maintenance stages
Where the product is applicable?
----------------
Proposal Summary
----------------
With this proposal we want to introduce a project aimed to enhance and simplify
operatinal maintenance of OpenStack cloud. Project provides service which uses
rule-based engine to inspect configurations of OpenStack
platform and find all kinds of architecture- and configuration-level glitches
and inconsistencies.
*# describe motivation behind rules
# describe rules reuse
# desribe rule-based inspection
# example rule
# mandatory rules vs. best-practice rules*
This engine will provide hints and best practices to increase reliability and
operational resilience of the cloud.
Rules engine
------------
------------------
Integration Points
------------------
TripleO integration points
--------------------------
.. image:: images/openstack_integration.png
This diagram shows integration points between OpenStack and config-validator:
* Source configuration data from **metadata services**
* Reuse **data collection methods** from TripleO program
* Create **configuration models** of all OpenStack services with default values of
the parameters and extended typization
* Use configuration files templates from TripleO program to define **actual
state** of configuration in configuration models
* Populate **architecture data model** with collected data
* Run **validator inspections** on the