Alexey, Dmitry,
We have executions counter and failed execution counter. If execution
succeeded only one counter is incremented, otherwise both counters are
incremented (some of execution failed semantic).
Why should we change this logic? These metrics can be interpreted in
different ways. I think this behaviour is correct. I think that you are
right also. But it seems that there is no significant difference in
treatment of these metrics.
On Tue, Jul 7, 2015 at 9:10 PM, Dmitriy Setrakyan <
[hidden email]>
wrote:
> On Tue, Jul 7, 2015 at 11:03 AM, Alexey Kuznetsov <
[hidden email]
> >
> wrote:
>
> > Igniters,
> >
> > I noticed that "IGNITE-1031 Query metrics updated incorrectly" was merged
> > into master.
> >
> > So, I tested it a little. And have a question.
> >
> > If I execute invalid sql (select from not existing table) I see that both
> > metrics updated: "Query Executions" and "Query Failures".
> >
> > I think this is NOT correct, why we increment "Query Executions" in case
> of
> > failed query?
> > I some how feel that this metric should be incremented only for
> > successfully executed queries.
> >
>
> Agree, current behavior seems to be wrong. Alexey K, can you please create
> a ticket for this?
>
>
> >
> > Thoughts?
> >
> > In case if metric updated correctly, then javadocs
> > for QueryMetrics.executions() should be more detailed.
> >
> > --
> > Alexey Kuznetsov
> > GridGain Systems
> > www.gridgain.com
> >
>
--
Andrey Gura
GridGain Systems, Inc.
www.gridgain.com
Locked
