[#172] Update HPA

[SilviaAmAm]

Fixes #172

Todo - tests

• Deploy OZ with autoscaling enabled for nginx/worker
• Deploy Objecten with autoscaling enabled for the workers

[alextreme]

Discussed:

• Silvia will make use of HPA via maykin-utils-lib for OpenArchiefBeheer via a separate PR
• For now we will only use maykin-utils-lib with OAB, this to ensure we don't run into dependancy problems while iterating on the shared library
• @Tiho0135 please verify the HPA changes made to the OpenZaak and Objecten charts before approving

[Tiho0135]

Yes, correct I did not verified the changes, will do that another day, apologies.

[SilviaAmAm]

Silvia will make use of HPA via maykin-utils-lib for OpenArchiefBeheer via a separate PR

#178

[Tiho0135]

I created Openzaak deployment with HPA enabled for all possibilities we offer, i.e. webapp, worker, nginx. All seems okay for the first two. For the nginx I do see reference to the nginx mertics and hence I do not know if the HPA in this case would work correctly. We use nginx k8s controller, and nginx ingress, so I do not know if our HPA is referring to nginx the same way it is referring to webapp and worker.
Not completely sure about, but wanted to share my thoughts.

[Tiho0135]

Using HPA with the following specifications:
targetCPUUtilizationPercentage: 80 targetMemoryUtilizationPercentage: 80
I get the following 7 new replicas pods for the worker deployment. Me personally I cannot explain the reason for this increase, since I do not add extra traffic myself. Maybe all is fine I just wanted to share this finding as well.

[Tiho0135]

I created Openzaak deployment with HPA enabled for all possibilities we offer, i.e. webapp, worker, nginx. All seems okay for the first two. For the nginx I do see reference to the nginx mertics and hence I do not know if the HPA in this case would work correctly. We use nginx k8s controller, and nginx ingress, so I do not know if our HPA is referring to nginx the same way it is referring to webapp and worker. Not completely sure about, but wanted to share my thoughts.

Actually, HPA for nginx in OpenZaak seems to be okay after some long time -> for example 2 hours. I would assume all is fine in that case.

[Tiho0135]

I think in general our HPA is configured correctly. When testing in our AKS cluster it seems the worker container is constantly autoscaling. In my opinion this is due to the k8s nodes low resources.
For this reason I will approve the changes.

[Tiho0135]

I creating PR for creating values with HPA settings for testing.

[alextreme]

@SilviaAmAm please resolve the conflicts, after that this PR can be merged