pg_profile 0.1.1
Major refactored pg_profile release. Due to total refactoring done in this release, migration from 0.0.7 is very difficult, so right now migration from 0.0.7 is not supported.
Installation
Provided file pg_profile-0.1.1.tar.gz is to be extracted into extension folder of postgresql installation:
sudo tar xzf pg_profile-0.1.1.tar.gz --directory $(pg_config --sharedir)/extension
You will need plpgsql and dblink extensions in your database, and pg_stat_statements extension if you need statements statistics. Then, you can create extension in postgresql database:
CREATE SCHEMA profile;
CREATE EXTENSION pg_profile SCHEMA profile;
Use provided pg_profile--0.1.1_sql.tar.gz file only for manual objects creation in restricted environments (when you have no file system access).
For further setup and usage details refer to doc/pg_profile.md file
Changelog
- Code reorganization, new Makefile
- Sequential scanned table now sorted by scanned pages estimation
- Order correction in top tables by read IO
- TOAST tables is now calculated and shown with main tables
- More fields in "Top SQL by I/O wait time" section
- Non-superuser install (see Privileges in doc)
- Collect postgres GUC parameters historic values and show them in reports
- Tablespaces support: size, growth and objects belong (Daria Vilkova)
- Fixed object rename issue
- New sections in report: "Top SQL by shared dirtied", "Top SQL by shared written", "Top tables by gets", "Top indexes by gets"
- Improved statement list in reports - now clicked statement_id is highlighted
- Workaround for amazon RDS
- Database sizes is now shown in reports (Daria Vilkova)
- Reports can be generated using time intervals (tstzrange type)
- snapshot_show() now displays info about stats reset, also report contains information about stats reset
- a lot of bugfixes and other improvements