Telegraf does not support a HorizontalPodAutoScaler

Telegraf currently does not support a HorizontalPodAutoScaler in its deployment and has to be pre-scalled to a number of replicas according to the peak event load it might have to process. This makes it difficult to dealing with a varying load dynamically and we are having to prescale it too high to mean peak load.

HPAs are standard for Kubernetes pods that can be horizontally scalled, which telegraf should be? See https://kubernetes.io/docs/tasks/run-application/horizontal-pod-autoscale/ for docs and https://github.com/bitnami/charts/blob/main/bitnami/nginx/templates/hpa.yaml for a reference example. We would be able to set 'autoscaling' values for min and max replicas, allowing it to scale when target cpu/mem resource utilization has reached on any pod and scale down as required when it has stabilized.

Please let us know if this is on the roadmap. If not would there be an objection to us submitting a PR? Thanks

Read the linked Kubernetes HPA documentation and Bitnami hpa.yaml example first, then inspect the Telegraf chart's deployment and values entry points. Done means the chart supports autoscaling values for minimum and maximum replicas and target CPU or memory utilization, with scaling up and down handled by Kubernetes.

Written by the indexing model from the issue text.