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
Currently Metaflow doesn't set the memory limit for jobs on Kubernetes. This can be an issue if there are two memory-hungry jobs running on the same node. If both flows together use more than is available on the node, then the node crashes with an OOM error which is difficult to debug.
I suggest that the memory limit be set to equal the memory request on Kubernetes to avoid this situation.
The text was updated successfully, but these errors were encountered:
Currently Metaflow doesn't set the memory limit for jobs on Kubernetes. This can be an issue if there are two memory-hungry jobs running on the same node. If both flows together use more than is available on the node, then the node crashes with an OOM error which is difficult to debug.
I suggest that the memory limit be set to equal the memory request on Kubernetes to avoid this situation.
The text was updated successfully, but these errors were encountered: