Database stat generation

[Bhanu, Yogesh]

Hello admins,
Is there any tool available inside Postgres to track what databases are being used and how . It will be nice to generate a per database stats (Inserts, Select, Drop) for a month or more and also track how the database grew big/small in the same time .

Cheers
yogesh

---------------------------(end of broadcast)---------------------------TIP 7: don't forget to increase your free space map settings

[Bill Montgomery]

Post by Bhanu, Yogesh
Hello admins,
Is there any tool available inside Postgres to track what databases are being used and how . It will be nice to generate a per database stats (Inserts, Select, Drop) for a month or more and also track how the database grew big/small in the same time .

Yogesh,

First, make sure you have set stats_start_collector and stats_row_level
to true. You may optionally set stats_command_string and
stats_block_level to true for more stats with a correspondingly greater
performance hit. Then, look at some of the pg_stat_* (and pg_statio_* if
you have block level stats turned on) tables/views.

A couple of the most useful tables (to me, anyhow): The pg_class view
has counts of pages and tuples per relation (tables are relkind 'r',
FYI). The pg_stat_database view shows the number of transactions
comitted/rolled back, which is useful to track rates of usage,
especially with a graphing tool like MRTG. To get the number of
inserts/updates/deletes on a table, check out the pg_stat_user_tables view.

For more info, check out the postgres docs on the subject:

http://www.postgresql.org/docs/7.3/interactive/monitoring-stats.html

Regards,

Bill Montgomery

---------------------------(end of broadcast)---------------------------
TIP 8: explain analyze is your friend