[
https://issues.apache.org/jira/browse/AMBARI-10348?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jonathan Hurley updated AMBARI-10348:
-------------------------------------
Attachment: AMBARI-10348.patch
> Alerts: Generate Server Side Alerts For Agent Health and Alert Staleness
> ------------------------------------------------------------------------
>
> Key: AMBARI-10348
> URL: https://issues.apache.org/jira/browse/AMBARI-10348
> Project: Ambari
> Issue Type: Bug
> Components: ambari-server
> Affects Versions: 2.0.0
> Reporter: Jonathan Hurley
> Assignee: Jonathan Hurley
> Priority: Critical
> Fix For: 2.1.0
>
> Attachments: AMBARI-10348.patch
>
>
> Due to alerts being run on distributed platforms - we have problem wherein if
> alert is scheduled to run on a host that goes down - its not obvious to
> customer that alerts are not running and something is wrong. We need to do 2
> thigns:
> 1. Generate an alert when not heard from Ambari Agent for quite sometime
> 2. Fire alerts which havent been run for quite sometime - saying they havent
> been run due to host not responding.
> Seems like 2 new alerts are required, both of which being "Server Side"
> (which is a new concept).
> - Ambari Server will need to maintain information about the last heartbeat
> from a host and produce alerts when a heartbeat has not been received. A new
> {{@AmbariService}} can handle this.
> - Ambari Server will need to periodically check the last timestamp of all
> enabled alert instances and determine if the alert has not run within a
> certain period of time. A new {{@AmbariService}} can handle this.
> We should utilize the alerts.json defined outside the stack since this
> affects hosts and alert instants and is not bound to a cluster.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
