I am using Keda http add-on to scale pods based on pending request. After installing it and setup forward request to keda-add-ons-http-interceptor-proxy service, I have simulated 502 requests to test scale pod automatically. But, number of pending request is equal or greater than spec.targetPendingRequests: 5, I see that HTTPScaledObject has restarted my old pods and then continuing to scale more pods to spec.replicas.max: 7.
I do not know this is bug or feature of http-add-on. Please helping me fix this ?
Expected Behavior
I expect http-add-on still keeps old pods and scales new pods basing on specification defined
Actual Behavior
My old pods is restarted by http-add-on when pending request is equal or greater spec.targetPendingRequests: 5
Steps to Reproduce the Problem
To simulate 502 requests
To watch pods by command watch kubectl -n <namesapce> get pod
When pending request is equal or greater than spec.targetPendingRequests: 5, I see that HTTPScaledObject restarted my old pods
Report
I am using Keda http add-on to scale pods based on pending request. After installing it and setup forward request to keda-add-ons-http-interceptor-proxy service, I have simulated 502 requests to test scale pod automatically. But, number of pending request is equal or greater than
spec.targetPendingRequests: 5
, I see thatHTTPScaledObject
has restarted my old pods and then continuing to scale more pods tospec.replicas.max: 7
.I do not know this is bug or feature of http-add-on. Please helping me fix this ?
Expected Behavior
I expect http-add-on still keeps old pods and scales new pods basing on specification defined
Actual Behavior
My old pods is restarted by http-add-on when pending request is equal or greater
spec.targetPendingRequests: 5
Steps to Reproduce the Problem
watch kubectl -n <namesapce> get pod
spec.targetPendingRequests: 5
, I see that HTTPScaledObject restarted my old podsLogs from KEDA HTTP operator
HTTP Add-on Version
0.7.0
Kubernetes Version
1.29
Platform
Other
Anything else?