Raising this question here as suggested in grafana/loki gitlab for opening question.
I am configuring our Loki to make it production ready for flow of read requests from users, dashborads and alerts. For this I am configuring autoscaling for loki-read component and setting:
autoscaling:
enabled: true
maxReplicas: 10
minReplicas: 4
as i can see in the the values doc it should be queryIndex.enabled: true but I have no references to this conf option anywhere. There is one closed request in github with this question but no attention to this part of question (or am I missed anything)?
As a result I do not see horizontalpodautoscaler created in the namespace for read, and looks like this config part is being ignored.
As I can see here it probably relies on loki.deployment.isScalable but also confused where it has to be set. Can someone please clarify.