Difference between revisions of "Entropy"
(→Summary) |
|||
Line 1: | Line 1: | ||
− | == Summary == | + | '''Revised on:''' {{REVISIONMONTH1}}/{{REVISIONDAY}}/{{REVISIONYEAR}} by {{REVISIONUSER}} |
+ | |||
+ | == Executive Summary == | ||
Entropy is a framework to write audit and repair scripts for openstack. It will allow writing cluster-check scripts, and define reactions to the errors/issues these bugs raise. | Entropy is a framework to write audit and repair scripts for openstack. It will allow writing cluster-check scripts, and define reactions to the errors/issues these bugs raise. | ||
Entropy will allow developers to write health checkers without worrying about deployment, setting up a Jenkins, integrating with an emailer, etc. It also allows definition of "reaction" scripts that wait on issues and take well defined actions (file a ticket, mark a hypervisor bad, etc). This automates reacting to failure at one level, and tackles inundating SEs with emails about (probably) minor issues. A potentially more important use is to aggregate failures, notice trends in failures, and developing a database of known failures to make dealing with new ones easier. | Entropy will allow developers to write health checkers without worrying about deployment, setting up a Jenkins, integrating with an emailer, etc. It also allows definition of "reaction" scripts that wait on issues and take well defined actions (file a ticket, mark a hypervisor bad, etc). This automates reacting to failure at one level, and tackles inundating SEs with emails about (probably) minor issues. A potentially more important use is to aggregate failures, notice trends in failures, and developing a database of known failures to make dealing with new ones easier. | ||
+ | |||
+ | ===Conceptual Example=== |
Revision as of 23:26, 9 December 2013
Revised on: 12/9/2013 by Harlowja
Executive Summary
Entropy is a framework to write audit and repair scripts for openstack. It will allow writing cluster-check scripts, and define reactions to the errors/issues these bugs raise.
Entropy will allow developers to write health checkers without worrying about deployment, setting up a Jenkins, integrating with an emailer, etc. It also allows definition of "reaction" scripts that wait on issues and take well defined actions (file a ticket, mark a hypervisor bad, etc). This automates reacting to failure at one level, and tackles inundating SEs with emails about (probably) minor issues. A potentially more important use is to aggregate failures, notice trends in failures, and developing a database of known failures to make dealing with new ones easier.