You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I read a bit about prefect work pools and workers. Here is some off the cuff information to save for later if I return to it:
.serve is a simple way to both update the prefect API with the latest deployments / flow definitions but also to start a process that waits for flow runs to be launched and runs them locally.
limitations in what it can do, and mostly seems to be a way to get started quickly
The level up is to separate the workers that actually run the jobs from where they are configed. You do this by creating a work pool which either your own workers listen to or more automatic workers that can run code on things like google cloud run listen to.
The work pool is managed as part of prefect cloud I think
It seems like the better long term way to define and work with prefect is through the cli prefect deploy and a yaml config file prefect.yaml.
Still overwhelming how many different ways there are to work with prefect, you really need your blinders on and ignore the "deployment blocks" "agents" etc etc stuff that doesn't matter and is just other ways to use prefect
The text was updated successfully, but these errors were encountered:
I read a bit about prefect work pools and workers. Here is some off the cuff information to save for later if I return to it:
.serve
is a simple way to both update the prefect API with the latest deployments / flow definitions but also to start a process that waits for flow runs to be launched and runs them locally.limitations in what it can do, and mostly seems to be a way to get started quickly
The level up is to separate the workers that actually run the jobs from where they are configed. You do this by creating a work pool which either your own workers listen to or more automatic workers that can run code on things like google cloud run listen to.
The work pool is managed as part of prefect cloud I think
It seems like the better long term way to define and work with prefect is through the cli
prefect deploy
and a yaml config fileprefect.yaml
.Still overwhelming how many different ways there are to work with prefect, you really need your blinders on and ignore the "deployment blocks" "agents" etc etc stuff that doesn't matter and is just other ways to use prefect
The text was updated successfully, but these errors were encountered: