Horizontal autoscaling for Operator #18
Description
As a user, I want my Lumigo operator to "just work" even when it's processing data from large clusters.
I would like a default horizontal pod autoscaler based on OpenTelemetry collector metrics from the telemetry-proxy container, as that is by far the component that will serve most load on large clusters.
Manual settings for replica size should be exposed as optional in Helm, but the overall experience should be an "auto" one.