The cloud-native orchestrator for the whole development lifecycle, with integrated lineage and observability.

dagster

Is there a way to launch a job run for every partition in a partition set at once for a non-time-based dynamcally generated partition config? And is it possible to do this on a schedule?

Hi <@U02KW95EFL3> - if you want to kick this off manually, you can launch a backfill from the "Partitions" tab of a run page

It's currently only possible to trigger one run at a time from a schedule. If you define a sensor, you can trigger multiple runs on the same tick.  E.g.
```@dynamic_partitioned_config(...)
def my_partitioned_config(...):
    ...

@sensor(job=...)
def my_sensor():
    if ...:
        for partition_key in my_partioned_config.get_partition_keys():
            yield RunRequest(tags={"partition": partition_key}, run_config=my_partitioned_config.get_run_config(partition_key))```


<@U011CET83FG> Any plans in the future to make scheduled backfills possible? Or is it a design decision for schedules to strictly be one run per tick?

<@U02KW95EFL3> - I created an issue for it: <https://github.com/dagster-io/dagster/issues/5714> (actually thought we had one already, but couldn't find it). I do think we'd like to enable this, though not sure on timeline

actually wait, I think this is already possible:

```@dynamic_partitioned_config(...)
def my_partitioned_config(...):
    ...

def evaluate_tick():
    if ...:
        for partition_key in my_partioned_config.get_partition_keys():
            yield RunRequest(tags={"partition": partition_key}, run_config=my_partitioned_config.get_run_config(partition_key))

my_schedule = ScheduleDefinition(job=..., execution_fn=evaluate_tick)```