You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Alexey Kuznetsov (JIRA)" <ji...@apache.org> on 2016/08/25 11:02:20 UTC

[jira] [Updated] (IGNITE-3769) Visor CMD: improve script execution in alert command

     [ https://issues.apache.org/jira/browse/IGNITE-3769?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Alexey Kuznetsov updated IGNITE-3769:
-------------------------------------
    Description: 
Current logic of executing custom scripts from alert command is not very usable and need to be improved.

Suppose alert could be in two states: "red" (alert conditions evaluated as true) and "green".

1) Alert checks condition each t milliseconds (already implemented).
2) If alert detected "red" state then custom script should be executed. If script is not available - retry with specified rate.
3) After alert detected "red" state it should continue to check condition each t milliseconds and execute script no more than throttling timeout "-i",
 this timeout could be "zero" that means that throttling disabled.
4) After alert detected "green" state then custom scrip executed with special "green state" flag.
5) When alert executed script it should pass to it how many times it was evaluated as "red state".

Looks good?

  was:Current logic of executing custom scripts from alert command is not very usable and need to be improved.


> Visor CMD: improve script execution in alert command
> ----------------------------------------------------
>
>                 Key: IGNITE-3769
>                 URL: https://issues.apache.org/jira/browse/IGNITE-3769
>             Project: Ignite
>          Issue Type: Task
>          Components: visor
>    Affects Versions: 1.7
>            Reporter: Alexey Kuznetsov
>             Fix For: 1.8
>
>
> Current logic of executing custom scripts from alert command is not very usable and need to be improved.
> Suppose alert could be in two states: "red" (alert conditions evaluated as true) and "green".
> 1) Alert checks condition each t milliseconds (already implemented).
> 2) If alert detected "red" state then custom script should be executed. If script is not available - retry with specified rate.
> 3) After alert detected "red" state it should continue to check condition each t milliseconds and execute script no more than throttling timeout "-i",
>  this timeout could be "zero" that means that throttling disabled.
> 4) After alert detected "green" state then custom scrip executed with special "green state" flag.
> 5) When alert executed script it should pass to it how many times it was evaluated as "red state".
> Looks good?



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)