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
From the next discussion #6334 (Editing ScaledObject on Workload in Production).
My proposal is to add a new (boolean) parameter at the fallback: useCurrentReplicasAsMinimum
If this parameter is set to true and it ends up in a fallback, at least the number of replicas that are running at that moment will continue to run. If the fallback.replicas is higher, this setting will be used.
Use-Case
When you are running a large number of pods and something goes wrong, Keda uses a fallback (if configured). However, the difference between the fallback.replicas and the current number of replicas can be large.
Something going wrong can also be temporary, which means it can suddenly scale down considerably and then return to the correct number of replicas. With this feature we ensure that this does not happen.
Is this a feature you are interested in implementing yourself?
Yes
Anything else?
If you have any objections or questions regarding this issue, please let me know :)
The text was updated successfully, but these errors were encountered:
Proposal
From the next discussion #6334 (Editing ScaledObject on Workload in Production).
My proposal is to add a new (boolean) parameter at the
fallback
:useCurrentReplicasAsMinimum
If this parameter is set to
true
and it ends up in a fallback, at least the number of replicas that are running at that moment will continue to run. If thefallback.replicas
is higher, this setting will be used.Use-Case
When you are running a large number of pods and something goes wrong, Keda uses a
fallback
(if configured). However, the difference between thefallback.replicas
and the current number of replicas can be large.Something going wrong can also be temporary, which means it can suddenly scale down considerably and then return to the correct number of replicas. With this feature we ensure that this does not happen.
Is this a feature you are interested in implementing yourself?
Yes
Anything else?
If you have any objections or questions regarding this issue, please let me know :)
The text was updated successfully, but these errors were encountered: