You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Briana Augenreich (Jira)" <ji...@apache.org> on 2022/08/25 14:30:00 UTC

[jira] [Updated] (HBASE-26809) Report client backoff time for server overloaded in ConnectionMetrics

     [ https://issues.apache.org/jira/browse/HBASE-26809?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Briana Augenreich updated HBASE-26809:
--------------------------------------
    Labels: patch-available  (was: )
    Status: Patch Available  (was: Open)

A master PR is available for review to enable metric reporting of an overloaded server backoff time. 

 

A PR containing the branch-2 changes (significantly more touch points) will follow after the master review. 

> Report client backoff time for server overloaded in ConnectionMetrics
> ---------------------------------------------------------------------
>
>                 Key: HBASE-26809
>                 URL: https://issues.apache.org/jira/browse/HBASE-26809
>             Project: HBase
>          Issue Type: New Feature
>            Reporter: Bryan Beaudreault
>            Assignee: Briana Augenreich
>            Priority: Major
>              Labels: patch-available
>
> Servers can throw CallQueueTooBigException and CallDroppedException when overloaded. As of HBASE-26807, these can have a configurable extra backoff to allow the server to recover. Depending on the server side queue implementation, different callers may be more or less impacted by server load. It is very useful to measure in the client how much time is spent backing off so we can see which clients are most affected by server overload.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)