My hobby plan database is currently over the 10,000 row limit, which is confusing to me because I have an hourly scheduled job that clears out old rows well in advance.
The psql-extras toolbelt reports this:
heroku pg:index_usage -a my-app
relname | percent_of_times_index_used | rows_in_table
----------------+-----------------------------+---------------
table_one | 26 | 8710
table_two | Insufficient data | 1235
table_three | 73 | 637
table_four | 0 | 429
table_five | | 1
(5 rows)
So that's over the limit... however when I SELECT COUNT(*) the rows in each table (which is what my scheduled job does), I can only find about 8,000 rows.
Take table_one, for example:
$ heroku psql -a my-app
--> Connecting to postgresql-my-db
psql (11.1, server 9.6.11)
SSL connection (protocol: TLSv1.2, ...)
Type "help" for help.
my-app::DATABASE=> SELECT COUNT(*) FROM table_one;
count
-------
5708
(1 row)
Does anyone know where these 3,000 extra rows are coming from?
Many thanks in advance.
An AUTOVACUUM seems to have sorted it out. Not sure why this happened after the thing has been running fine for over a year, but never mind.