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
This is a followup to #3397 where our node pools got a node-purpose = core|notebook|dask-worker label added to them unconditionally. The unconditional change will force our node pools to upgrade, which often leads to disruption. Due to this, the change hasn't yet been applied to all our clusters node pools.
Here is how I think we should make a hard-to-do-all-at-once infra change collaboratively over time:
at one point in time: introduce a temporarily opt-out change and opting out of the change systematically
incrementally over time and by all engineers: incrementally stop opting out of the change across our clusters
cleanup the temp opt-out config when its not longer referenced
This is a followup to #3397 where our node pools got a
node-purpose = core|notebook|dask-worker
label added to them unconditionally. The unconditional change will force our node pools to upgrade, which often leads to disruption. Due to this, the change hasn't yet been applied to all our clusters node pools.Here is how I think we should make a hard-to-do-all-at-once infra change collaboratively over time:
Action points
terraform, gcp: add opt-out var to collaboratively and opportunistically apply node-purpose infra labels #3406
Clusters with partial upgrades remaining currently
The text was updated successfully, but these errors were encountered: