[
https://issues.apache.org/jira/browse/HDFS-14406?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16820537#comment-16820537
]
Erik Krogen commented on HDFS-14406:
------------------------------------
Thanks for the ping, [~elgoiri]. I do have some comments:
* Creating a new {{rpcuser}} metrics context seems potentially overkill. This
is still providing details about RPC calls, so it seems to me that it can still
live under the {{rpcdetailed}} context with the sub-name "rpcuser".
* {{RpcUserMetrics}} has an {{init()}} method that initializes it with the
methods of a {{protocol}}. This is completely meaningless in this context,
since {{RpcUserMetrics}} does not store metrics about method names. This should
be removed.
* {{RpcUserMetrics}} and {{RpcDetailedMetrics}} are nearly identical, can we
have a common base class which they both inherit from, or even just let
{{RpcUserMetrics extends RpcDetailedMetrics}}? The only differences are the
init step, and the name which is passed to the registry.
* In the Javadoc for {{IPC_SERVER_USER_METRICS_DEFAULT}}, can we use a
{{@link}} tag to refer to the key?
* In the description within {{core-site.xml}}, the sentence has a typo:
{quote}
Set the whether we enable per user rpc processing time metrics on namenode rpc
server.
{quote}
should be
{quote}
Set whether we enable per user rpc processing time metrics on NameNode rpc
server.
{quote}
(An extra "the", and also IMO NameNode should be capitalized)
* Within {{Metrics.md}}, I think the descriptions could use updating:
{quote}
| *username*`NumOps` | Total number of the times RPC with username is called |
| *username*`AvgTime` | Average turn around time of the user RPC in
milliseconds |
{quote}
What does "turn around time" even mean? I would prefer:
{quote}
| *username*`NumOps` | Total number of times RPCs were called by username |
| *username*`AvgTime` | Average processing time of the user's RPCs in
milliseconds |
{quote}
> Add per user RPC Processing time
> --------------------------------
>
> Key: HDFS-14406
> URL: https://issues.apache.org/jira/browse/HDFS-14406
> Project: Hadoop HDFS
> Issue Type: Improvement
> Affects Versions: 3.2.0
> Reporter: Xue Liu
> Assignee: Xue Liu
> Priority: Minor
> Fix For: 3.2.0
>
> Attachments: HDFS-14406.001.patch, HDFS-14406.002.patch,
> HDFS-14406.003.patch, HDFS-14406.004.patch
>
>
> For a shared cluster we would want to separate users' resources, as well as
> having our metrics reflecting on the usage, latency, etc, for each user.
> This JIRA aims to add per user RPC processing time metrics and expose it via
> JMX.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
