[jira] [Commented] (HBASE-17448) Export metrics from RecoverableZooKeeper

["Andrew Purtell (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HBASE-17448?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16382493#comment-16382493 ] 

Andrew Purtell commented on HBASE-17448:
----------------------------------------

Discussion on HBASE-19769

[~elserj]
{quote}
bq. I feel this should come out in 1.5 since it is already out in 2.0, at least.
Yeah, I'd agree with you.
bq. Do you know if this an issue in 1.4x [~elserj] also sir?
I haven't explicitly tested it, but I would assume the issue also exists there. Looking at DynamicMetricsRegistry, I see we're referencing MetricsInfoImpl which was the problematic class being accessed across these ClassLoader boundaries.
bq. Although this shipped in 1.4.0, 1.4.1, and 1.4.2, I have no problem pulling it from 1.5 if problematic, or even from 1.4.3 if more than problematic approaching broken. Please advise.
Yeah, pulling from 1.4 would probably be best also (I'm not sure why I didn't ask the question then, apologies). I think there are two potential fixes:
 # For branch-1, lift anything package-private from metrics2 we have to rely on into hbase itself
 # For branch-2, use the hbase-metrics API and avoid metrics2 completely (I think helping HBASE-18409 get in would be helpful, maybe necessary)
{quote}

[~stack]
{quote}
Arghhhh... Should have kept an eye on HBASE-18409 .... Thats a shame. Its too late now... but hadoop metrics are only ever going to be pain producing.
{quote}

Myself:
{quote}
Ok. Let me revert the commit of HBASE-17448 on branch-1 and branch-1.4, referencing this JIRA, and state that the plan is to eventually reimplement with hbase-metrics-api or after HBASE-18409 gets in 
{quote}

> Export metrics from RecoverableZooKeeper
> ----------------------------------------
>
>                 Key: HBASE-17448
>                 URL: https://issues.apache.org/jira/browse/HBASE-17448
>             Project: HBase
>          Issue Type: Improvement
>          Components: Zookeeper
>    Affects Versions: 1.3.1
>            Reporter: Andrew Purtell
>            Assignee: Chinmay Kulkarni
>            Priority: Major
>              Labels: patch
>             Fix For: 2.0.0, 1.4.0
>
>         Attachments: HBASE-17448-branch-1.patch, HBASE-17448.patch, HBASE-17448.patch
>
>
> Consider adding instrumentation to RecoverableZooKeeper that exposes metrics on the performance and health of the embedded ZooKeeper client: latency histograms for each op type, number of reconnections, number of ops where a reconnection was necessary to proceed, number of failed ops due to CONNECTIONLOSS, number of failed ops due to SESSIONEXIPRED, number of failed ops due to OPERATIONTIMEOUT. 
> RecoverableZooKeeper is a class in hbase-client so we can hook up the new metrics to both client- and server-side metrics reporters. Probably this metrics source should be a process singleton. 



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)