Skip to main content

Configure default maintenance settings

The Settings page holds the defaults for all maintenance jobs. A job uses these values unless you change them on the job itself.

Janitor starts with safe defaults. You can use it without changing anything here.

Prerequisites​

  • Your role includes maintenance:manage. To only view the settings, maintenance:read is enough.

Open the settings​

  1. Select Janitor → Settings in the sidebar.
  2. Find the card you want to change. There are four cards: Schedule, Maintenance window, Concurrency, and Maintenance options.
  3. Select the edit icon in the top right corner of the card.
  4. Change the values, then select Save.

Settings page with the Schedule, Maintenance window, Concurrency, and Maintenance options cards

Schedule​

FieldWhat it setsAllowed values
ScheduleThe default time for jobs to run: every day, every week on a weekday, or every month on a day, at a time you pick.—
TimezoneThe time zone of the schedule.A time zone from the list
Schedule checker (s)How often Janitor checks for jobs that are due, in seconds.10 to 3600
Query timeout (s)How long Janitor waits for a response from a maintenance query, in seconds.30 to 3600

Maintenance window​

The maintenance window is the time of day when job schedules are allowed, for example 00:00 to 06:00. Pick a time when the Lakehouse is quiet.

FieldWhat it sets
StartThe start of the window, HH:MM.
EndThe end of the window, HH:MM.

The window can go past midnight, for example 22:00 to 06:00.

When a window is set, you cannot save a job whose next run falls outside it. Janitor shows an error and asks you to pick another time. When no window is set, any schedule is allowed.

Concurrency​

These limits control how many runs work at the same time. Higher values finish maintenance sooner but put more load on the Lakehouse.

FieldWhat it setsAllowed values
Max concurrent runsRuns that can work at the same time in the whole Lakehouse. 1 by default.1 to 100
Max concurrent runs per catalogRuns that can work at the same time in one catalog.1 to the value of Max concurrent runs
Max manual runsRuns started with Run that can work at the same time.1 to the value of Max concurrent runs
Max concurrents startScheduled runs Janitor can start in one check. 1 by default.1 to the value of Max concurrent runs

For example, when Max concurrent runs is 5, the other three limits can be 1 to 5.

Maintenance options​

These options are the defaults for new jobs. You can change them for each job later.

OptionOn by defaultRetention
OPTIMIZEYes—
EXPIRE SNAPSHOTSYesSnapshot retention (days), 7 by default
ORPHAN CLEANUPNoOrphan retention (days), 7 by default

Retention can be 7 to 365 days.

warning

ORPHAN CLEANUP removes files for good. It cannot be undone.

Keep Orphan retention (days) the same as or longer than Snapshot retention (days). If it is shorter, Janitor saves your change but shows a warning: files that a kept snapshot still uses could be removed.

See What maintenance does.

If something goes wrong​

SituationWhat it meansWhat to do
A value is out of rangeThe number is below the minimum or above the maximum.Enter a value in the allowed range. See Input rules.
A limit is higher than Max concurrent runsMax concurrent runs per catalog, Max manual runs, and Max concurrents start cannot be higher than Max concurrent runs.Lower the value, or raise Max concurrent runs first.
Retention is less than 7 daysRetention must be at least 7 days.Pick 7 days or more.
Save fails with an access errorYour role has only maintenance:read.Ask your administrator for maintenance:manage.

Next steps​