+sandy

up

up

Back to the drawing table on this one

remove subtract in favor of replace

up

looks fine, maybe remove the patch on warnings?

up

end in tests should be the end of the partition time, add additional tests to test for when end is not specified, update get_schedule_range_partitions logic to be more clear

This example needs to show up in the docs somewhere, probably in https://docs.dagster.io/deployment/run-coordinator. Similar to https://docs.dagster.io/deployment/guides/docker#example, we should highlight the fact that a custom run coordinator can be used, and that an example is a run coordinator that injects run tags.

let's kick off some more runs just to see how this holds up against flakiness. was honestly expecting more of an performance gain - current tests are running at around 20-23m and this went to 19m. not a bad win, but will probably need to apply this to all tox.ini files

test 5

test 4

test 3

test 2

test 1

We should also pass in -n since we're using pytest-xdist https://github.com/pytest-dev/pytest-xdist#speed-up-test-runs-by-sending-tests-to-multiple-cpus? looks like we're only using this in dagit tests right now

Will this example be explicitly mentioned in the docs somewhere?

The better fix would be to fix the resolver for canTerminate, which involves changing the logic in some run launchers

rebase

rebase

ignore airflow tests failing, haven't rebased on master yet

up

rebase

+chris feedback

In D8099#212491, @cdecarolis wrote:

In the general case of someone missing a resource however, I'm not sure what the benefit of displaying keys is, however.

In D8070#211857, @dgibson wrote:

Not sure what our back-compat preservation strategy is for the python client, particularly since it's so new... is it OK to swap out the underlyint query for reloadRepositoryLocation here (and just delete the old snapshot from the python client backcompat tests)?

rebase

black