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
We recently upgraded our GKE Autopilot cluster to 1.29 and to our surprise, it is no longer calculating the resource allocation the same way, and the default metaflow resources are now "just a little" too large, resulting in our workflows being rejected.
I found the defaults are configured here, and we also have many ways to configure the resources, but it would be ideal for us to be able to configure the default resources via an environment variable, much the same way we can set the nodeSelector.
The text was updated successfully, but these errors were encountered:
We recently upgraded our GKE Autopilot cluster to 1.29 and to our surprise, it is no longer calculating the resource allocation the same way, and the default metaflow resources are now "just a little" too large, resulting in our workflows being rejected.
I found the defaults are configured here, and we also have many ways to configure the resources, but it would be ideal for us to be able to configure the default resources via an environment variable, much the same way we can set the nodeSelector.
The text was updated successfully, but these errors were encountered: