[ 
https://issues.apache.org/jira/browse/SENTRY-1872?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16117944#comment-16117944
 ] 

Alexander Kolbasov commented on SENTRY-1872:
--------------------------------------------

There are two problems here:

1) You are adding 500ms latency to every request even in the case where 
everything is working just fine.
2) How do you know that 500ms is enough time to detect a bad server?

> Check if a Sentry server is accessible before connecting to it
> --------------------------------------------------------------
>
>                 Key: SENTRY-1872
>                 URL: https://issues.apache.org/jira/browse/SENTRY-1872
>             Project: Sentry
>          Issue Type: Bug
>          Components: Sentry
>    Affects Versions: 2.0.0
>            Reporter: Arjun Mishra
>            Assignee: Arjun Mishra
>         Attachments: SENTRY-1872.01.patch
>
>
> For Sentry HA, if one of the server is unreachable for any reason, we would 
> still try to connect to it, and only switch to other server after a timeout 
> exception is thrown. We should instead check if the server is reachable first 
> before we try connecting to it



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Reply via email to