Autoscaling with kubernetes v1.13 HPA objects not being created.

[lumjjb]

Is this a BUG REPORT or FEATURE REQUEST?:

Bug report.

What happened:

Upon creating an autoscale rule for a function in kubeless, kubeless autoscale ls did not show any results. In addition, no HPA objects were found in the cluster.

What you expected to happen:

A rule for autoscaling should be shown in kubeless autoscale ls together with its associated hpa objection.

How to reproduce it (as minimally and precisely as possible):

On a kubernetes v1.13 cluster:

kubeless function deploy hello --runtime python2.7 --from-file test.py
kubeless autoscale create hello --metric qps --value 1000
kubeless autoscale ls

Anything else we need to know?:

Logs from kubectl logs -n kubeless -l kubeless=controller -c kubeless-function-controller

time="2019-01-16T19:11:31Z" level=error msg="Function can not be created/updated: the server could not find the requested resource (post servicemonitors.monitoring.coreos.com)" pkg=function-controller
time="2019-01-16T19:11:31Z" level=error msg="Error processing default/hello (will retry): the server could not find the requested resource (post servicemonitors.monitoring.coreos.com)" pkg=function-controller
time="2019-01-16T19:11:31Z" level=info msg="Processing change to Function default/hello" pkg=function-controller
time="2019-01-16T19:11:33Z" level=error msg="Function can not be created/updated: the server could not find the requested resource (post servicemonitors.monitoring.coreos.com)" pkg=function-controller
time="2019-01-16T19:11:33Z" level=error msg="Error processing default/hello (giving up): the server could not find the requested resource (post servicemonitors.monitoring.coreos.com)" pkg=function-controller
ERROR: logging before flag.Parse: E0116 19:11:33.097796       1 function_controller.go:185] the server could not find the requested resource (post servicemonitors.monitoring.coreos.com)
ERROR: logging before flag.Parse: W0117 05:33:35.463334       1 reflector.go:341] github.com/kubeless/kubeless/pkg/controller/function_controller.go:139: watch of *v1beta1.Function ended with: too old resource version: 12797 (61331)
ERROR: logging before flag.Parse: W0119 12:25:54.604742       1 reflector.go:341] github.com/kubeless/kubeless/pkg/controller/function_controller.go:139: watch of *v1beta1.Function ended with: too old resource version: 61331 (318089)
ERROR: logging before flag.Parse: W0120 23:16:45.441150       1 reflector.go:341] github.com/kubeless/kubeless/pkg/controller/function_controller.go:139: watch of *v1beta1.Function ended with: too old resource version: 318089 (481123)
ERROR: logging before flag.Parse: W0122 10:08:45.655731       1 reflector.go:341] github.com/kubeless/kubeless/pkg/controller/function_controller.go:139: watch of *v1beta1.Function ended with: too old resource version: 481123 (644265)

Environment:

• Kubernetes version (use kubectl version):

Client Version: version.Info{Major:"1", Minor:"13", GitVersion:"v1.13.2", GitCommit:"cff46ab41ff0bb44d8584413b598ad8360ec1def", GitTreeState:"clean", BuildDate:"2019-01-10T23:35:51Z", GoVersion:"go1.11.4", Compiler:"gc", Platform:"linux/amd64"}
Server Version: version.Info{Major:"1", Minor:"13", GitVersion:"v1.13.2", GitCommit:"cff46ab41ff0bb44d8584413b598ad8360ec1def", GitTreeState:"clean", BuildDate:"2019-01-10T23:28:14Z", GoVersion:"go1.11.4", Compiler:"gc", Platform:"linux/amd64"}

• Kubeless version (use kubeless version):
  Kubeless version: v1.0.1

• Cloud provider or physical cluster:
  Kubeadm cluster on GCE

[andresmgot]

Hi @lumjjb,

Thank you for the report. It seems that we need to update the monitoring dependencies that we are using to support autoscaling in 1.13. Just to double check, the deployment of the function work as expected and you can trigger functions but you cannot add autoscaling, isn't it?

[lumjjb]

Yup! That is right, function deployment and calling works perfectly!

[cscetbon]

Any news on that issue ?

[andresmgot]

We plan to update the supported versions for Kubeless @cscetbon but we haven't had the time to do so yet. Stay tuned!

If in the meantime someone wants to pick this and contribute an update I will be happy to help.

[chalocodekare]

Is this issue still not fixed ? Is there any workaround ?

[andresmgot]

This is still an issue, but it seems that it may be related to the way the hpa are listed since in #1219 the resource is being created. In any case, accepting contributions here if anyone wants to take a look.