For those using airflow to schedule meltano (separ...
# best-practices
d
For those using airflow to schedule meltano (separate airflow from meltano maintained), do you have meltano in a separate repo / docker from airflow?
wondering how to handle dependency conflicts between airflow, meltano, and potentially dbt
s
We have everything in the same docker project 😄 we run dbt & airflow through meltano, so we don't really have dependency issues
d
Ahh yeah we cannot unfortunately so hoping to figure out what best practice is when they are separate My main question would be, if they’re in separate containers, how do you trigger the sync - probably through the API?
s
Trigger a sync as in trigger a run?
(Of airflow)
d
Trigger a meltano sync
Since they’d be in unrelated environments - i’d imagine we’d have to call the meltano API since we couldn’t just use the CLI
or perhaps a dockeroperator
s
Yeah that could probably work 😄 I have never had this issue, so sadly I can't speak to it. I'm fairly confident though that it is common place in the community to have a separate airflow integration from your meltano one, so someone should be able to help you out! Sorry for not being more helpful
l
I have deployed meltano along with Airflow in the same docker compose project but as different docker compose services that are in the same docker network. The Airflow worker communicates with meltano using ssh. I tuned the
meltano.py
dag generator to use
SSHOperator
instead of
BashOperator