Scheduled Tasks (Cron Job)

Register a scheduled task: hologres.hg_register_cron_job

Register a new cron job with hologres.hg_register_cron_job. Syntax:

SELECT hologres.hg_register_cron_job(
  job_name => '<job_name>',
  command => '<command>',
  [comment => '<comment>',]
  [schedule => '<schedule>',]
  [warehouse_id => '<warehouse_id>',]
  [db_name => '<db_name>',]
  [user_name => '<user_name>',]
  [retry_count => '<retry_count>',]
  [starts_time => '<starts_time>',]
  [ends_time => '<ends_time>',]
  [interval_text => '<interval_text>']
);

Parameter details are listed in Parameters.

Modify a scheduled task: hologres.hg_alter_cron_job

Modify a registered cron job with hologres.hg_alter_cron_job. Syntax:

SELECT hologres.hg_alter_cron_job(
  job_name => '<job_name>',
  command => '<command>',
  comment => '<comment>',
  schedule => '<schedule>',
  warehouse_id => '<warehouse_id>',
  db_name => '<db_name>',
  retry_count => '<retry_count>',
  active => '<active>',
  starts_time => '<starts_time>',
  ends_time => '<ends_time>',
  interval_text => '<interval_text>'
);

The active parameter is unique to hologres.hg_alter_cron_job. Set it to true to enable the job, or false to pause scheduling while retaining the registration. Other parameters match those in Register a scheduled task.

Delete a scheduled task: hologres.hg_unregister_cron_job

Delete a cron job with hologres.hg_unregister_cron_job. Syntax:

SELECT hologres.hg_unregister_cron_job('<job_name>');

View scheduled tasks: hg_user_cron_jobs / hg_user_cron_tasks

Query registered cron jobs (not yet executed) with the hologres.hg_user_cron_jobs view:

SELECT * FROM hologres.hg_user_cron_jobs;

Query execution history and errors with the hologres.hg_user_cron_tasks view:

SELECT * FROM hologres.hg_user_cron_tasks;

Example 1: Run a data import every 5 minutes

Import data from table a to table b every 5 minutes. You can use either schedule or interval_text.

Method 1: Using schedule (cron expression)

SELECT hologres.hg_register_cron_job(
  job_name => 'job_import_data',
  command => 'INSERT INTO b SELECT * FROM a;',
  schedule => '0 */5 * * * ?'
);

Method 2: Using interval_text

SELECT hologres.hg_register_cron_job(
  job_name => 'job_import_data',
  command => 'INSERT INTO b SELECT * FROM a;',
  interval_text => '5 minutes'
);

Example 2: Run TRUNCATE every hour within a specified time range

During April 2026, run TRUNCATE on table tbl in the test_db database at the top of every hour, using resources on warehouse 1 and retrying up to 5 times on failure.

SELECT hologres.hg_register_cron_job(
  job_name => 'job_truncate_tbl',
  comment => 'Scheduled TRUNCATE',
  schedule => '0 0 */1 * * ?',
  command => $$TRUNCATE tbl$$,
  warehouse_id => 1,
  db_name => 'test_db',
  retry_count => 5,
  starts_time => '2026-04-01 00:00:00+08',
  ends_time => '2026-04-30 23:59:59+08'
);

How do I troubleshoot a scheduled task that fails to run?

1. Query the hologres.hg_user_cron_tasks view for execution records and error details.

2. Check whether the SQL statement has syntax errors.

3. Check whether the user specified by user_name has the required permissions.

What should I do if a scheduled task misses a scheduling cycle?

When concurrency resources are exhausted or a task overruns its next scheduled time, take the following steps:

1. Adjust the scheduling interval to prevent task pile-up.

2. Optimize SQL performance to reduce per-execution runtime.

3. Evaluate the concurrent task volume to stay within the default 100 concurrency limit.

How do I pause a scheduled task?

Set active to false in hologres.hg_alter_cron_job to pause a cron job. Set active to true to resume it.

SELECT hologres.hg_alter_cron_job(
  job_name => 'job_import_data',
  command => 'INSERT INTO b SELECT * FROM a;',
  comment => 'Data import task',
  schedule => '0 */5 * * * ?',
  warehouse_id => -1,
  db_name => 'mydb',
  retry_count => 0,
  active => false,
  starts_time => NULL,
  ends_time => NULL,
  interval_text => NULL
);