You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@sentry.apache.org by "Na Li (JIRA)" <ji...@apache.org> on 2018/05/29 21:19:00 UTC
[jira] [Commented] (SENTRY-2250) Optimize the logic of fetching the
notification_log entries
[ https://issues.apache.org/jira/browse/SENTRY-2250?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16494289#comment-16494289 ]
Na Li commented on SENTRY-2250:
-------------------------------
Are we OK to fetch notification for each DDL task that needs sync? If so, then when getting sync event, sentry server can start a thread from a thread pool and fetch the notification events. If there are max number of multiple pending threads, we should use periodical fetch
> Optimize the logic of fetching the notification_log entries
> -----------------------------------------------------------
>
> Key: SENTRY-2250
> URL: https://issues.apache.org/jira/browse/SENTRY-2250
> Project: Sentry
> Issue Type: Improvement
> Components: Sentry
> Reporter: kalyan kumar kalvagadda
> Assignee: kalyan kumar kalvagadda
> Priority: Major
>
> Currently sentry fetches entries from notification_log in HMS periodically. This could result in delay 0.5 sec to 1 sec in fetching the events and processing them.
> As HMS threads are blocked till sentry processes the notification events, this delay in processing notifications would result in delay in query executions.
>
> Proposal is to proactively fetch the notification log entires from HMS.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)