Closed corentin59 closed 10 months ago
Not really sure to understand your use cases ? You still have the database and storage that will be stateful (even if you could delegate to others resource in the cloud) The flow definition are stored on that database and we advocate for using the CI/CD to deploy.
We have some detail about the architecture you want?
In fact, we currently see Kestra as a tooling component in our Kubernetes cluster, which is configured during the initialization of a new cluster. The lifespan of a Kubernetes cluster in our environment can be quite limited, ranging from a few days to a few months. With this in mind, it would be nice to have Kestra initialized on the fly with generic flows ready for use. Users can then add their own flows afterward. One could envision creating these generic flows as a sidekick or through API calls... However, if HELM supports it, it might be simpler for us at the moment ?
We don't provide this in the helm charts but you could emulate this with a kubernetes job I think. You could:
./kestra flow namespace update --server http://<service-url> <your.namespace> <dir> --no-delete
Does this will work for you?
THANKS,
That should do !
Corentin
Closing since the solution seems ok for you, reopen if not
Feature description
Hello,
I am currently using Kestra via Kubernetes with Helm deployment. I would like to maintain a fairly stateless installation in the sense that my requirements are currently limited (meaning without storage). Is there a way during Helm deployment to create flows using the values.yaml file ?
Thx