configuration problems about pgobserver HOT 6 CLOSED

Hi,

Thanks for the feedback! The Python error was indeed a bug, I fixed it now in master (b0140d4).

About the other issues - all those "Perf. *" screens not showing data is probably OK, as they're only meant to show possible problems as sudden rise in sequential scans etc, which might not be the case.

For sprocs info not being shown - you should 1st check the "sproc_schemas_monitoring_configuration" table and adjust it if needed. Here is decided from which schemas information is actually gathered. By default for example the "public" schema is left out (as using it is not a good practice).

For Pg_Stat_Statements report - there the gatherer probably is by default not instructed to collect this data (you could check the monitor_data.stat_statements_data table) as out of box Postgres doesn't have this information. You can fix this from "/hosts" screen by updating the Setting column key "statStatementsGatherInterval" to some > 0 value.

from pgobserver.

Hey,

thanks for you quick reply. You are right, i missed the configuration of statStatementsGatherInterval for the Pg_Stat_Statements report and already configured that accordingly later on. It works now as expected.

Regarding the "Top Tables" page bug, i checked out your update. And this also works now. Thanks for that.

Another problem i had when using you vagrant configuration is with the Dockerfile of the gatherer. The Dockerfile of the gatherer tries to add pgobserver-gatherer-1.0.4-jar-with-dependencies.jar, while it mus be pgobserver-gatherer-1.0.5-jar-with-dependencies.jar, note the different version number.

Nice work that you put into this project.

from pgobserver.

Having a similar problem over here. SELECT * FROM pg_stat_statements; does return rows, but the monitor_data.stat_database_data table stays empty, and so is the /perfstatstatements page. I already set the "statStatementsGatherInterval": 5, setting on the /hosts table.

@kmoppel Any ideas?

These are my postgresql.conf settings:

shared_preload_libraries = 'pg_stat_statements'         # (change requires restart)

pg_stat_statements.max = 10000
pg_stat_statements.track = all

from pgobserver.

Hi Erik,
If I understand you right you expect data in stat_database_data? Then you should set statDatabaseGatherInterval to a value bigger than 0. statStatementsGatherInterval indeed gathers only pg_stat_statements. For this it is required to have zz_utils.get_stat_statements installed on the monitored cluster..

from pgobserver.

My bad, I meant stat_statements_data, I copy/pasted the wrong table name.

Anyways, thank you for your response. I double checked the get_stat_statements.sql and I figured out it didn't do anything after executing it, even though it displayed a success message.
The problem was, the extension still wasn't enabled on the database, so after executing the query CREATE EXTENSION pg_stat_statements; and executing the .sql file again, it worked 👍 Thanks again.

from pgobserver.

Thanks for the feedback again @ErikvdVen ! Btw, if you see that something could be improved in the "setup" section of the readme file, feel free to create a PR :) Closing the issue for now.

from pgobserver.