pg_cron debugging guide

Last edited: 2/21/2025

This is a general guide for debugging pg_cron. Below lists issues and how to debug them

Cannot create/edit/delete cron jobs

Cron jobs can only be modified with the respective SQL functions:

If you are trying to make changes, use the cron functions. If the cron functions are inaccessible, contact Support



Cron Jobs are not running

You should consider initiating a software upgrade in the Infrastructure Settings if your Postgres version is below v15.6.1.122. Upgrading will give you access to pg_cron v1.6.4+, which has many bug fixes and auto-revive capabilities.

Debugging steps:

Check to see if "pg_cron scheduler" is active

pg_cron operates as the pg_cron scheduler process within Postgres. Use the below query to check if the worker is active

1
2
3
4
5
6
7
8
9
10
SELECT pid as process_id, usename as database_user, application_name, backend_start as when_process_began, wait_event_type, state, query, backend_typeFROM pg_stat_activity where application_name ilike 'pg_cron scheduler';

If the query does not return a row, the worker has died. To revive it, you must go to the General Settings and initiate a fast reboot:

Screenshot 2024-10-29 at 12 27 24 AM

Check the cron.job_run_details table for more information

pg_cron creates logs in its own table cron.job_run_details". The below query checks for issues from the past 5 days :

1
2
3
4
5
6
7
8
SELECT *FROM cron.job_run_detailsWHERE (status <> 'succeeded' AND status <> 'running') AND start_time > NOW() - INTERVAL '5 days'ORDER BY start_time DESCLIMIT 10;

Respond to the errors exposed appropriately.

Long running jobs may show timeout errors. For jobs that are intended to execute for extended periods, consider wrapping their queries in functions with custom timeouts (guide).


Check if there are too many cron jobs running concurrently

pg_cron supports up to 32 concurrent jobs, each using a database connection. If too many jobs are running simultaneously, space them out to prevent connection overload and job failure.

The below queries shows all pg_cron jobs:

1
2
3
4
5
-- All jobsselect schedule, jobname, command from cron.job;-- Count jobsselect COUNT(*) from cron.job;

The below query can be used to find actively querying jobs:

1
2
3
4
5
6
7
8
9
10
SELECT pid as process_id, usename as database_user, application_name, backend_start as when_process_began, wait_event_type, state, query, backend_typeFROM pg_stat_activity where application_name ilike 'pg_cron';

You can view your concurrent peak connection usage throughout the day at the bottom of the Reports Dashboard


Check for database strain

Unfortunately, excessive resource strain can slow down or disrupt jobs.

Go to the reports page (or Supabase Grafana if you have it setup), and check for signs of resource exhaustion. If it's clear your database is under pressure, consider upgrading your compute add-on or following the advice from one of the optimization guides:

It is important to make sure you are running the latest release of pg_cron (1.6.4) if you're noticing strain. It is the most robust.


Check the log explorer for more information

Although pg*cron records errors in the cron.job_run_details table, in rare cases, more information can be found in the general Postgres logs. You can check the Log Explorer for failure events with the following query

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
select cast(postgres_logs.timestamp as datetime) as timestamp, event_message, parsed.error_severity, parsed.user_name, parsed.query, parsed.detail, parsed.hint, parsed.sql_state_code, parsed.backend_type, parsed.application_namefrom postgres_logs cross join unnest(metadata) as metadata cross join unnest(metadata.parsed) as parsedwhere regexp_contains(parsed.error_severity, 'ERROR|FATAL|PANIC') and regexp_contains(parsed.application_name, 'pg_cron')order by timestamp desclimit 100;

If you're interested in modifying the query, there is an advanced guide for navigating the Postgres logs and a general purpose one for applying filters.


Create custom logs within cron jobs

If it's still not clear what is occurring you may be able to capture more logs by running the pg_cron query inside a database function:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
create or replace function log_example()returns voidlanguage plpgsqlas $$begin -- Logging start of function raise log 'logging start of cron function call: (%)', (select now()); -- INSERT LOGIC HERE -- Logging end of function raise log 'logging end of cron function call: (%)', (select now()); exception -- Handle exceptions here if needed when others then raise exception 'An error occurred in cron function <insert name here>. ERROR MESSAGE: %', sqlerrm;end;$$;

You can then search for your custom messages in the Logs Interface


Upgrading pg_cron version

The current version of pg*cron on Supabase is 1.6.4. It comes with a few bug fixes. You should consider upgrading to Postgres v15.6.1.122+ in the Infrastructure Settings to get the latest extension.


Contacting support and the maintainers

Although Supabase includes the extension, it is maintained by Citus (a Microsoft subsidiary). You can contact Support for more help, but you should also consider creating an issue in the pg_cron repo.