dip CLI; from the Pipelines page you watch it, pause or resume its schedule, run it on demand, and
read every run’s results. Runs execute on your data plane — Dippr keeps the run history and the per-model cost.
A pipeline’s schedule is part of its code. Add both owner=owner(service_user="<service user>") and
schedule=schedule(cron="0 6 * * *", timezone="UTC") to the manifest(...) in pipeline.py (import
owner and schedule from altimate_pipeline). A schedule needs an owner, because a scheduled run acts
as it. Then publish with dip pipeline push and attach the owner’s key with dip pipeline set-owner.
Without a schedule the pipeline is manual-only. There is no command or console control that sets a
cron.
The cron has five fields (minute, hour, day of month, month, day of week), each within its range, and the
timezone is an IANA name such as Europe/Brussels or UTC. dip pipeline check looks only at the
shape, so a cron such as 61 * * * * or a timezone such as Mars/Olympus passes it. dip pipeline push then refuses the pipeline with an error that names the value, and nothing is registered.
A scheduled run starts at its cron time, in the schedule’s timezone, and never before it. The scheduler
checks for due runs about once a minute, so a run usually starts within a minute after its cron time:
0 6 * * * runs shortly after 06:00. A run can start later when the scheduler is delayed, or when a
scheduled run could not start and is waiting to try again.
dip pipeline push publishes the image to your workspace’s registry on the data plane, and runs pull it
from there. A pipeline can only use an image from that registry that your own workspace pushed:
registering one from another workspace’s namespace is refused. An image on any other registry is used as
you give it.
Availability
Pipelines are included in the standard tenant bundle through thepipelines entitlement. A platform administrator can disable it for a
tenant. In a tenant without it:
- The Pipelines page says “Pipelines are not included in this workspace”.
- Every
dip pipelinecommand that reaches Dippr stops with a message that the workspace’s plan does not include the feature.dip pipeline pushstops before it builds or pushes an image. The commands that only work on your machine (new,check,build,sdk-sync) keep working. dipleavesdip pipelineout of its help, anddip pipeline newdoes not write the pipeline agent skill.- Scheduled runs do not start. Pipelines you already registered are kept, and their schedules resume when the entitlement is added back.
The pipelines list
Each row is one pipeline. The columns:
Type in the search box to narrow the rows on screen by name. Use Load more at the bottom to pull in
the next page.
Status
The Status column shows one of these:
A Blocked badge appears next to the status when the scheduler will not run the pipeline even though
it is scheduled and not paused. Hover it for the reason, for example “no owner credential”. A blocked
pipeline needs that reason resolved before its schedule fires.
When the scheduler cannot start a pipeline’s scheduled run, the row shows how many scheduled runs in a
row could not be started, for example “2 failed starts”. After each failed start the scheduler waits
before trying again: at least 1 minute after the first, doubling each time, up to 30 minutes. The wait
is a minimum, so a retry can come later than that, but never sooner. The count returns to 0 as soon as a scheduled run starts, or when the scheduler finds
the pipeline already running.
A run that starts and then fails, for example because a dbt model fails, does not add to this count
and does not delay the schedule. The next scheduled run starts on time, so a fix you push is picked up
at the next slot. To see failed runs, check the run status or the pipeline’s
run history.
Run a pipeline now
Select Run now on a row to start a run immediately. The run moves through Queued then Running on its own; open the pipeline to watch it and read its results.Pause and resume
Select Pause to stop a pipeline’s schedule. A run already under way is not affected — pausing only stops the next scheduled run from starting. Select Resume to turn the schedule back on. The Next run column reflects the change. The first run after you resume is at the schedule’s first cron time after you resume. Cron times that passed while the pipeline was paused are not made up. Running a pipeline, and pausing or resuming it, need permission to change pipelines. Without it these buttons are disabled and their tooltip names the access you need.Next steps
Pipeline owner
Set up the service user a pipeline runs as, and rotate or revoke its key.
A pipeline's runs and results
Open a pipeline to see its declaration, its run history, and everything a run produced.
Query engines
Pipelines run on the engines you register here.