You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "Jason Dere (JIRA)" <ji...@apache.org> on 2016/11/09 23:07:58 UTC
[jira] [Updated] (HIVE-15149) Add additional information to ATSHook
for Tez UI
[ https://issues.apache.org/jira/browse/HIVE-15149?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Jason Dere updated HIVE-15149:
------------------------------
Attachment: HIVE-15149.1.patch
Work in progress, added the following fields to the ATS event:
Hive query name
Hive configs
HiveServer2 IP address
Client IP
execution mode (mr/tez/llap/spark)
Hive instance type (cli/hs2)
Tables read/written
Fixed thread name (originally was ATSHook thread)
> Add additional information to ATSHook for Tez UI
> ------------------------------------------------
>
> Key: HIVE-15149
> URL: https://issues.apache.org/jira/browse/HIVE-15149
> Project: Hive
> Issue Type: Improvement
> Reporter: Jason Dere
> Assignee: Jason Dere
> Attachments: HIVE-15149.1.patch
>
>
> Additional query details wanted for TEZ-3530. The additional details discussed include the following:
> Publish the following info ( in addition to existing bits published today):
> Application Id to which the query was submitted (primary filter)
> DAG Id (primary filter)
> Hive query name (primary filter)
> Hive Configs (everything a set command would provide except for sensitive credential info)
> Potentially publish source of config i.e. set in hive query script vs hive-site.xml, etc.
> Which HiveServer2 the query was submitted to
> *Which IP/host the query was submitted from - not sure what filter support will be available.
> Which execution mode the query is running in (primary filter)
> What submission mode was used (cli/beeline/jdbc, etc)
> User info ( running as, actual end user, etc) - not sure if already present
> Perf logger events. The data published should be able to create a timeline view of the query i.e. actual submission time, query compile timestamps, execution timestamps, post-exec data moves, etc.
> Explain plan with enough details for visualizing.
> Databases and tables being queried (primary filter)
> Yarn queue info (primary filter)
> Caller context (primary filter)
> Original source i.e. submitter
> Thread info in HS2 if needed ( I believe Vikram may have added this earlier )
> Query time taken (with filter support )
> Additional context info e.g. llap instance name and appId if required.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)