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
I recently came across and documented the following bug in which the Github Runner Scaler can cause a panic with a simple typo in a parameter in the scaled job manifest. This panic put the KEDA operator itself into a crashloop. It would be good to add some protection in the call flow to capture panics introduced by a scaler and avoid impacting the operator as a whole.
Wrap the call to the scaler such that a panic within the scaler code is trapped, logged accordingly and the operator continues executing any other scaling activities on the cluster.
Is this a feature you are interested in implementing yourself?
No
Anything else?
No response
The text was updated successfully, but these errors were encountered:
Proposal
I recently came across and documented the following bug in which the Github Runner Scaler can cause a panic with a simple typo in a parameter in the scaled job manifest. This panic put the KEDA operator itself into a crashloop. It would be good to add some protection in the call flow to capture panics introduced by a scaler and avoid impacting the operator as a whole.
#6156
Use-Case
Wrap the call to the scaler such that a panic within the scaler code is trapped, logged accordingly and the operator continues executing any other scaling activities on the cluster.
Is this a feature you are interested in implementing yourself?
No
Anything else?
No response
The text was updated successfully, but these errors were encountered: