Open davenquinn opened 1 year ago
The logging here makes clear we are just accomplishing a fair number of tasks (cluster version checking, etc.) before we start running the command in the container. Maybe we should have a whitelist of commands that we run right off the bat without configuring the entire environment? And/or we could have some sort of cache of instance information (this is needed anyway to pass configuration details to the frontend admin).
We can also shift to the architectural model we are beginning to adopt for Macrostrat, where the assumption is that management scripts will mostly be run as non-containerized processes.
We are having a hard time running the
sparrow create-user
command because it seems that a lot of code is being loaded on each app run. This may have something to do with database migrations.