reason for default PGSTAT_ACTIVITY_SIZE

From: frank joerdens <fiskadoro(at)gmail(dot)com>
To: pgsql-general(at)postgresql(dot)org
Cc: dev <dev(at)woome(dot)com>
Subject: reason for default PGSTAT_ACTIVITY_SIZE
Date: 2010-01-05 17:02:53
Message-ID: 7d10d2df1001050902p639f4046o87473984ef9a3770@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

From my angle, it looks as if the default for PGSTAT_ACTIVITY_SIZE is
too small, or rather that it ought to be configurable at least, so
that longer current_query strings that are shown via
pg_stat_get_backend_activity() in the system view pg_stat_activity
aren't truncated.

The use case is to watch pg_stat_activity when troubleshooting a
busy/struggling live server which is building up a queue of long
running queries, to be able to see which of your SQL is most
problematic, and where it's coming from: We append a code reference as
a comment to all SQL that is sent by the ORM (we use a python web
application development framework, namely Django), in order to figure
out where it originates, which is otherwise tricky with SQL that is
generated by a translation layer - and this code reference is often
cut off because the string is limited.

Is there a good reason to set PGSTAT_ACTIVITY_SIZE at 256 and is my
only option to recompile the server? Is there a practical
limit/drawback to making the variable say 4 or 8 times the default?

I am referring to this post here:

http://archives.postgresql.org/pgsql-sql/2002-05/msg00087.php

and this one:

http://archives.postgresql.org/pgsql-novice/2004-02/msg00187.php

according to which recompiling would be the only option, but that was
a few years back so maybe it's a different story now.

Regards,

Frank

Responses

Browse pgsql-general by date

  From Date Subject
Next Message John R Pierce 2010-01-05 17:16:22 Re: Large tables, ORDER BY and sequence/index scans
Previous Message sabrina miller 2010-01-05 17:00:51 Re: Using table name in column for joining