I an new to Postgres (9.3 on Ubuntu 14.04) monitoring and I have been using Zabbix/PG_monz.
I am seeing errors that my database has "too many temp bytes" and that my database is "too large". My actual backup files are 400 MB daily.
I am wondering if anyone is familiar with PG_Monz or just what this error might be about. Thank you in advance!
1 Answer 1
I found the answer with regards to how and why this is happening. PG_Monz is using the following query to track temp bytes:
UserParameter=psql.db_temp_bytes[*],psql -h 1ドル -p 2ドル -U 3ドル -d 4ドル -t -c "select temp_bytes from pg_stat_database where datname = '5ドル'"
However, this select statement does not show the current temp bytes, instead it shows an increment of all temp bytes ever created since the last stats reset. IE: It's not showing a delta over a time period which would be the accurate number.
Looking at this query and then checking the folder, I see it's empty and therefore my server actually has no temp bytes.
select current_setting('data_directory') || '/base/pgsql_tmp'
Source: https://www.marshut.net/wnsmt/postgresql-9-2-4-temp-files-never-released.html
The same goes for a DB being too large:
Zabbix threshhold is incorrect for the way PostgreSQL reports size of databases:
See info about select pg_database_size(dbname)