[jira] [Commented] (SENTRY-1115) Add caching to avoid huge performance hit

["Gregory Chanan (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SENTRY-1115?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15256793#comment-15256793 ] 

Gregory Chanan commented on SENTRY-1115:
----------------------------------------

I'd like this in Solr as well.  One question I'd have is if we'd really even need an LRU cache?  Is the number of objects so large in the kafka case that we'd ever need to evict something?  I can't see that happening in the Solr case.

> Add caching to avoid huge performance hit
> -----------------------------------------
>
>                 Key: SENTRY-1115
>                 URL: https://issues.apache.org/jira/browse/SENTRY-1115
>             Project: Sentry
>          Issue Type: Sub-task
>            Reporter: Ashish K Singh
>            Assignee: Ashish K Singh
>             Fix For: 1.8.0
>
>
> Kafka is highly performant and serves huge number of requests per second. This performance will be severely hurt, if all these requests start hitting Sentry server. To keep the performance hit minimal, it is required to cache authorize requests and corresponding results. LRUMap can be used with a configurable size.



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