[
https://issues.apache.org/jira/browse/CASSANDRA-7962?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jonathan Shook updated CASSANDRA-7962:
--------------------------------------
Description:
Count the usages of prepared statements on the server side, and expose them via
JMX. The original query structure should be included. Provide a nodetool
command to see the counts. Expose a meaningful id which can be used in other
places, such as with JMX clients. If enumerations are used to identify these
for easier discussion, enumerate according to the statement id.
Allow for "since last access" deltas, or "since server start", and require this
parameter to be specified. When the deltas form is used, show the number of
seconds represent by the current data.
This would allow easier identification of access pattern distribution in the
case that prepared statements are being used. It would provide useful metrics
for diagnosis, testing, and monitoring.
nodetool command syntax:
nodetool ( spusagecountsummary | spusagecountdetails ) ( deltas | alltime ) [
ks [ table ] ]
Example nodetool outputs:
nodetool spusagecountsummary deltas appks usertable
(count, statement id), since 234233 seconds ago
56304 24ad327f9bb2578de663fc92336703dd
5 8743b52063cd84097a65d1633f5c74f5
1 663fc92336703dd24ad327f9bb2578de
0 92336703dd24ad327f9bb2578de663fc
0 bb2578de663fc922336703dd24ad327f9
nodetool spusagecountdetails alltime appks usertable
(count, statement id,\n query\n\n)
56320304 24ad327f9bb2578de663fc92336703dd
select user from usertable where userid=?;
265 8743b52063cd84097a65d1633f5c74f5
insert into usertable (userid,...) ...
11 663fc92336703dd24ad327f9bb2578de
select tags from tagcloud where userid=?;
... and so forth ...
was:
Count the usages of prepared statements on the server side, and expose them via
JMX. The original query structure should be included. Expose a meaningful id
which can be used in other places, such as with JMX clients. If enumerations
are used to identify these for easier discussion, enumerate according to the
statement id.
Allow for "since last access" deltas, or "since server start", and require this
parameter to be specified. When the deltas form is used, show the number of
seconds represent by the current data.
This would allow easier identification of access pattern distribution in the
case that prepared statements are being used. It would provide useful metrics
for diagnosis, testing, and monitoring.
nodetool command syntax:
nodetool ( spusagecountsummary | spusagecountdetails ) ( deltas | alltime ) [
ks [ table ] ]
Example nodetool outputs:
nodetool spusagecountsummary deltas appks usertable
(count, statement id), since 234233 seconds ago
56304 24ad327f9bb2578de663fc92336703dd
5 8743b52063cd84097a65d1633f5c74f5
1 663fc92336703dd24ad327f9bb2578de
0 92336703dd24ad327f9bb2578de663fc
0 bb2578de663fc922336703dd24ad327f9
nodetool spusagecountdetails alltime appks usertable
(count, statement id,\n query\n\n)
56320304 24ad327f9bb2578de663fc92336703dd
select user from usertable where userid=?;
265 8743b52063cd84097a65d1633f5c74f5
insert into usertable (userid,...) ...
11 663fc92336703dd24ad327f9bb2578de
select tags from tagcloud where userid=?;
... and so forth ...
> usage counters on prepared statements, summary and details
> ----------------------------------------------------------
>
> Key: CASSANDRA-7962
> URL: https://issues.apache.org/jira/browse/CASSANDRA-7962
> Project: Cassandra
> Issue Type: New Feature
> Components: Core
> Reporter: Jonathan Shook
> Priority: Minor
>
> Count the usages of prepared statements on the server side, and expose them
> via JMX. The original query structure should be included. Provide a nodetool
> command to see the counts. Expose a meaningful id which can be used in other
> places, such as with JMX clients. If enumerations are used to identify these
> for easier discussion, enumerate according to the statement id.
> Allow for "since last access" deltas, or "since server start", and require
> this parameter to be specified. When the deltas form is used, show the number
> of seconds represent by the current data.
> This would allow easier identification of access pattern distribution in the
> case that prepared statements are being used. It would provide useful metrics
> for diagnosis, testing, and monitoring.
> nodetool command syntax:
> nodetool ( spusagecountsummary | spusagecountdetails ) ( deltas | alltime ) [
> ks [ table ] ]
> Example nodetool outputs:
> nodetool spusagecountsummary deltas appks usertable
> (count, statement id), since 234233 seconds ago
> 56304 24ad327f9bb2578de663fc92336703dd
> 5 8743b52063cd84097a65d1633f5c74f5
> 1 663fc92336703dd24ad327f9bb2578de
> 0 92336703dd24ad327f9bb2578de663fc
> 0 bb2578de663fc922336703dd24ad327f9
> nodetool spusagecountdetails alltime appks usertable
> (count, statement id,\n query\n\n)
> 56320304 24ad327f9bb2578de663fc92336703dd
> select user from usertable where userid=?;
> 265 8743b52063cd84097a65d1633f5c74f5
> insert into usertable (userid,...) ...
> 11 663fc92336703dd24ad327f9bb2578de
> select tags from tagcloud where userid=?;
> ... and so forth ...
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
