Re: New SQL counter statistics view (pg_stat_sql)

From: Haribabu Kommi <kommi(dot)haribabu(at)gmail(dot)com>
To: Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>
Cc: Peter Eisentraut <peter(dot)eisentraut(at)2ndquadrant(dot)com>, Robert Haas <robertmhaas(at)gmail(dot)com>, Andres Freund <andres(at)anarazel(dot)de>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: New SQL counter statistics view (pg_stat_sql)
Date: 2016-10-12 03:21:25
Message-ID: CAJrrPGc44YAjgDY4mnpVuk9WLB6y-+oj7gkqFZKahK_sfQM0ag@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Thu, Sep 29, 2016 at 3:45 PM, Haribabu Kommi <kommi(dot)haribabu(at)gmail(dot)com>
wrote:

>
>
> On Thu, Sep 22, 2016 at 3:05 AM, Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>
> wrote:
>
>> Peter Eisentraut wrote:
>>
>> > How about having the tag not be a column name but a row entry. So you'd
>> > do something like
>> >
>> > SELECT * FROM pg_stat_sql WHERE tag = 'ALTER VIEW';
>> >
>> > That way, we don't have to keep updating (and re-debating) this when new
>> > command types or subtypes are added. And queries written for future
>> > versions will not fail when run against old servers.
>>
>> Yeah, good idea.
>>
>
> Yes, Having it as a row entry is good.
>
>
>
>> Let's also discuss the interface from the stats collector. Currently we
>> have some 20 new SQL functions, all alike, each loading the whole data
>> and returning a single counter, and then the view invokes each function
>> separately. That doesn't seem great to me. How about having a single C
>> function that returns the whole thing as a SRF instead, and the view is
>> just a single function invocation -- something like pg_lock_status
>> filling pg_locks in one go.
>>
>> Another consideration is that the present patch lumps together all ALTER
>> cases in a single counter. This isn't great, but at the same time we
>> don't want to bloat the stat files by having hundreds of counters per
>> database, do we?
>
>
> Currently, The SQL stats is a fixed size counter to track the all the
> ALTER
> cases as single counter. So while sending the stats from the backend to
> stats collector at the end of the transaction, the cost is same, because of
> it's fixed size. This approach adds overhead to send and read the stats
> is minimal.
>
> With the following approach, I feel it is possible to support the counter
> at
> command tag level.
>
> Add a Global and local Hash to keep track of the counters by using the
> command tag as the key, this hash table increases dynamically whenever
> a new type of SQL command gets executed. The Local Hash data is passed
> to stats collector whenever the transaction gets committed.
>
> The problem I am thinking is that, Sending data from Hash and populating
> the Hash from stats file for all the command tags adds some overhead.
>

I tried changing the pg_stat_sql into row mode and ran the regress suite to
add different type of SQL commands to the view and ran the pgbench test
on my laptop to find out any performance impact with this patch.

HEAD PATCH
pgbench - select 828 816

Here I attached the pg_stat_sql patch to keep track of all SQL commands
based on the commandTag and their counts. I attached the result of this
view that is run on the database after "make installcheck" just for
reference.

Regards,
Hari Babu
Fujitsu Australia

Attachment Content-Type Size
pg_stat_sql_output.txt text/plain 6.1 KB
pg_stat_sql_row_mode_1.patch application/octet-stream 25.3 KB

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Pavel Stehule 2016-10-12 03:52:19 Re: Remove "Source Code" column from \df+ ?
Previous Message Thomas Munro 2016-10-12 03:19:44 Re: DISTINCT with btree skip scan